Sensors serve as an extension of human senses, facilitating the heightened perception and profound comprehension of the world. Among various kinds of sensors, magnetic field sensors play an increasingly vital role across diverse areas, such as biomedicine, aerospace, military, and industry, as well as fundamental scientific research. Currently, the state-of-the-art magnetic sensors are based on Superconducting Quantum Interference Devices (SQUIDs), with exceptional sensitivity at the fT/Hz^1/2 level. However, their reliance on cryogenic systems leads to high operational costs and limited applications. In recent decades, significant advancements have been made in other magnetic sensors that can work at room temperature, including Hall sensors, optical atomic magnetometers, diamond Nitrogen Vacancy (NV) center magnetic sensors, etc. This review sheds light on an emerging, high-sensitivity magnetic field sensor known as the Optical Microcavity Magnetic sensor (OMM). OMMs utilizing high-quality factor optical microcavities have enabled precision sensing of magnetic fields, featuring high sensitivity, broad bandwidth, and low power consumption. In this review, we focus on recent advances using three types of OMMs: magnetostrictive magnetic sensors, torque magnetic sensors, and magneto-optic magnetic sensors. We provide an overview of the magnetic field sensing mechanisms employed by these three OMMs and survey the recent relevant progress. Furthermore, we discuss the sensitivity improvement methods through response enhancement and noise suppression. Finally, this review provides an outlook on the potential of OMMs for magnetic induction tomography and corona current monitoring.Based on their sensing mechanisms, OMMs can be typically categorized into magnetostrictive magnetic sensors, torque magnetic sensors, and magneto-optic magnetic sensors. The magnetostrictive magnetic sensors combine magnetostrictive materials with optical microcavities. The applied magnetic field induces strain within the magnetostrictive material, which can drive the mechanical motion of the microcavity, leading to a change in the radius of the microcavity. This radius change shifts the optical resonance, resulting in a periodic change in the intracavity field. Consequently, the magnetic field can be optically read out using the intensity-modulating mechanism, with sensitivity enhanced by the dual-resonance of optical and mechanical modes. Initially, researchers achieved OMMS by embedding magnetostrictive particles such as Tefernol-D into optical microcavities. The sensitivity has been improved by optimizing the geometric parameters and improving the overlap between the magnetostrictive material and the microcavity. In efforts to achieve scalable and stable magnetostrictive magnetic sensors, researchers have sputtered the magnetostrictive thin films onto the optical microcavity and even integrated it with a waveguide. Recently, sensitivities of 26 pT/Hz^1/2 and 585 pT/Hz^1/2 have been achieved by using the magnetostrictive materials of Terfenol-D particles and Terfenol-D thin films, respectively.Torque magnetic sensors consist of a torque mechanical resonator coated with a magnetic material layer and an optical microcavity such as a photonic crystal microcavity or an optical whispering gallery mode microcavity, with an air gap between the mechanical resonator and the optical cavity. The magnetic field exerts a moment on the torque, causing excitation of the torsional mode and altering the distance between the optical microcavity and the torque oscillator. This change in distance affects the effective refractive index of the optical mode, resulting in periodic shifts in the resonance frequency of the optical mode. Therefore, the magnetic field or the magnetic moment of the material can be optically read out using the intensity-modulating mechanism. Magneto-optic magnetic sensors involve injecting the magnetic fluid into microcavities such as capillary and Fabry-Perot fiber cavities. The magnetic nanoparticles within the magnetic fluid can form chain-like structures as a magnetic field is applied, resulting in a change in the refractive index of the magnetic fluid, which induces a shift in the optical resonance. Thus, the magnetic field can be read out by the variation of the optical transmission.Sensitivity is a critical parameter for magnetic sensors as it determines their ability to detect weak magnetic fields. In the case of magnetic-optic magnetic sensors, sensitivity is commonly evaluated by measuring the mode shift caused by a unit magnetic field, expressed in units of nm/mT. On the other hand, for optomechanical magnetic sensors like magnetostrictive magnetic sensors and torque magnetic sensors, sensitivity is typically characterized by the noise-equivalent magnetic field. This metric represents the amplitude of the magnetic field that the sensor can detect at a signal-to-noise ratio of 1, with a measurement resolution bandwidth of 1 Hz, measured in T/Hz^1/2. The noise sources of optomechanical magnetic sensors consist of thermal noise from the nonzero-temperature bath, shot noise from the laser, and back action noise originating from the radiation pressure shot noise heating. The sensitivity of optomechanical magnetic sensors is minimal near the mechanical resonance frequency where thermal noise dominates, and degrades in the shot-noise-limited regime. In this review, we focus on the optomechanical magnetic sensors as an example to discuss the sensitivity improvement methods through signal enhancement and noise suppression. In the regime of response enhancement, sensitivity can be improved by selecting materials that exhibit higher sensitivity to the magnetic field and optimizing the structural design of the magnetic sensor. Additionally, the response to the magnetic field can be amplified by employing a flux concentrator which enhances the magnetic field signal. In the regime of noise suppression, the sensitivity can be improved by reducing the shot noise level of the laser using squeezed light or performing joint measurements with entangled light.While there is still progress to be made in improving the sensitivity, stability, and scalability of OMMs, their unique advantages make them a promising technology for future magnetic field sensing applications. Continued research and development efforts are expected to address the current challenges and unlock the full potential of OMMs in practical application scenarios, such as magnetic induction tomography and corona current detection.