Optical skyrmions, as an emergent cutting-edge topic in optics and photonics, extend the concept of non-singular topological defects to topological photonics, providing extra degrees of freedom for light–matter interaction manipulations, optical metrologies, optical communications, etc^{[1]}. The realization of artificial optical skyrmions did not occur until 2018^{[2,3]}, while the starting point of pursuits of optical skyrmions could date to Maxwellian and Kelvin’s era, as shown in Fig. 1. The history of the rheological of skyrmions concept is somewhat similar to the homeward journey of the Greek myths hero Odysseus with twists and turns. The story goes all the way back to the old days when scientists discovered electromagnetism. Inspired by the fact of the curl field nature of magnetism, Maxwell believed that electromagnetism should have a rotational origin and proposed a model of ether vortices to derive the equations of electromagnetism^{[4]}. After that, Lord Kelvin went a step further to propose an atomic model based on knots of swirling ethereal vortices immersed in an ether sea^{[5]}. In the 1870s, there were huge debates over Kelvin’s vortex atoms model. Maxwell, a vortex atom enthusiast, promoted the model in his influential Encyclopedia Britannica article, “Atom.” The opponent, like Boltzmann, said that the model lacks any proof of the validity of the equations. Along with the discovery of electrons and nuclei, the vortex atom hypothesis was finally abandoned, whereas the attractive features of those knots, including discreteness and immutability, have never been forgotten, and the idea of knots and knot invariants spawned a key modern physics conception, topological defects in the field theory. Around 60 years later, shown in Fig. 1, the general interests of physicists changed from atoms to sub-atoms. The idea of knots returned to the stage, and it was employed by Skyrme to describe the nuclei^{[6,7]}. In Skyrme’s picture, protons and neutrons are depicted as topological knot defects excitation in the three-component pion field, well known as skyrmions. The number of knot twists, or knot invariants, is equal to the number of nucleons in the nuclei. And by skyrmions, certain nucleus states had also correctly been predicted. Moreover, different from Kelvin’s vortex atom hypothesis, the skyrmions in the nuclei are based on the nonlinear field theory with pion-pion interactions. And the nonlinear interaction physically guarantees that the skyrmions are stable under perturbations, in addition to the topological reason. Although it is accepted that the skyrmion is historically the first example of a topological defects model, as the saying goes, the course of true love never did run smoothly. Along with the discovery of quarks, the skyrmion model was overlooked. Unexpected turns were associated with the rise of condensed matter physics. In condensed matter physics, a collectively large number of atoms and electrons with fruitful symmetries, interactions, as well as phases offers a platform to effectively construct various topological defect excitations, for example, vortices in superconductors, monopoles in spin ices, and skyrmions in non-centrosymmetric magnetic systems^{[8–10]}. Most importantly, condensed matter systems are sensitive to diversified external fields, leading to steerable manipulations of those topological defects with both versatility and precision. Therefore, since entering the condensed matter physics era, the concept of topology has been not only used to explain natural matter but also to control or even design matter, being the core of modern physics and related disciplines. Nevertheless, the skyrmion in condensed matter physics is still not a trouble-free journey. In the 1960s and 70s, it was believed that skyrmions were not expected to exist in most condensed matter systems due to the Hobart-Derrick theorem^{[9,10]}. Even in one of the most well-known papers of Nobel prize winners Kosterlitz and Thouless, it is said that: “If we regard the direction of magnetization in space as giving a mapping of the space on to the surface of a unit sphere (actually it is exactly skyrmions), this invariant (skyrmions number) measures the number of times the map of the space encloses the sphere. This invariant is of no significance in statistical mechanics”^{[11]}. However, as marked in Fig. 1, in 1989 A.N. Bogdanov and collaborators^{[12,13]} uncovered that magnetic materials with a broken inversion symmetry or, in other words, so-called non-centrosymmetric magnetic systems, could support magnetic skyrmions. And it took another 20 years to realize the magnetic skyrmions experimentally^{[14,15]}. Since then, magnetic skyrmions have become one of the hottest topics in condensed matter physics, constituting a promising new direction for data storage and spintronics^{[10,16]}. Of course, it is not the end of the Odyssey journey of skyrmions. Ahead, there are still lots of challenges and opportunities, such as simultaneously increasing the stability and the transition temperature of the magnetic skyrmions with nanometer size and realizations of skyrmions in other classical systems such as light. Recently, the storyline went to optical skyrmions, as shown in Fig. 1.