The tactile perception of a flexible humanoid robotic finger is essential for advancing applications in the fields of intelligent robotics, human-computer interaction, and prosthetics. This work presents a novel multi-parameter sensing method using fiber Bragg grating embedded in a cantilevered robotic finger structure. It aims to address the challenges posed by strain and temperature cross-sensitivity in tactile sensing systems. By embedding a differential sensing array of fiber Bragg grating sensors into the robotic finger, the design achieves accurate, real-time measurement of tactile force, contact temperature, and contact position while maintaining robustness and flexibility.The robotic finger is designed to closely replicate the structural and functional characteristics of a human finger, providing both dexterity and adaptability. The cantilevered finger bone is fabricated from polylactic acid through 3D printing, ensuring lightweight and high-strength properties. Six fiber Bragg grating sensors are symmetrically embedded within the cantilevered structure in pairs to form a differential sensing array. This differential configuration allows for the temperature compensation by sensing the same temperature and the opposite strain. This design effectively decouples strain and temperature effects, significantly improving the reliability of the sensor readings under varying environmental conditions.To ensure high precision, each fiber Bragg grating sensor in the array was carefully calibrated. Calibration experiments were performed to determine the stress and temperature sensitivities of the fiber Bragg grating sensors. The experimental results revealed that the embedded fiber Bragg grating array achieved a stress sensitivity of 112.898 pm/N and a temperature sensitivity of 81.185 pm/℃. Furthermore, the sensors demonstrated excellent accuracy and stability across a wide temperature range (10~50 ℃), with a minimum average measurement error of -0.025 6 nm. These results confirm the effectiveness of the differential sensing array in distinguishing between the effects of strain and temperature, making the system suitable for the tactile sensing tasks in practical applications.In addition to force and temperature sensing, the robotic finger can accurately detect the contacting position. By extracting the standard deviation features of the central wavelength shifts from the fiber Bragg grating sensors, we employed a support vector machine algorithm to classify different contact locations. The classification model was trained on data collected from multiple contact points across the finger's surface. The dataset is divided into a 70% training set (105 samples) and a 30% test set (45 samples) to ensure that the model's generalization ability. The support vector machine-based model achieved a contact position classification accuracy of 95.5%, demonstrating the system's capability to perform precise contact detection. This high level of accuracy is critical for applications that require fine manipulation, such as robotic surgery, where precise tactile feedback ensures safe and effective operation, and advanced prosthetic devices, where accurate sensory information enhances the user experience.In conclusion, this work demonstrates that fiber Bragg grating sensing technology, when integrated into a flexible robotic finger structure, provides a highly effective solution for multi-parameter tactile sensing. The system can accurately measure force, temperature, and contact position, featuring a lightweight design and high flexibility. This technology holds broad potential applications, particularly in scenarios where precise and reliable tactile feedback is essential. These include not only robotic surgery and advanced prosthetics but also collaborative robotics and human-computer interaction systems. Future research will focus on refining the design of the sensor array, optimizing the algorithmic approaches for data processing, and exploring additional applications in environments that demand precise tactile sensing.