This study proposes an intelligent optimization algorithm inspired by astrophysical kinematics, called gravitational-collision optimization algorithm (G-COA). In a galaxy, the mass of the star accounts for more than 90% of the whole galaxy. According to the size of gravity, it is in the absolute dominant position of the star system. Several planets revolve around them in different orbits. In addition, most of the planets are surrounded by several satellites orbiting the planets. Satellites often encounter the collision of small celestial bodies, resulting in changes in their speed and position, or getting rid of the gravity of the central celestial body, or falling into the central celestial body. The optimization algorithm proposed in this study is based on the motion behavior of these celestial bodies to update the solution, and then find the optimal solution after multiple iterations. Each celestial body represents a feasible solution. G-COA can dynamically adjust exploration and development functions at different stages of iteration. Based on some existing astrophysics-inspired optimization algorithms, this study improves the gravitational search strategy and integrates more celestial motion behaviors into the algorithm. In order to verify the effectiveness of each search strategy, ablation experiments are carried out. Based on the CEC2022 test function set, several latest algorithms and highly cited algorithms are compared. The optimization results of 12 of the 16 test functions are optimal. Finally, two engineering constraint problems are listed, which further proves that G-COA is a promising algorithm.