In order to solve the problem of inaccuracy of path point planning of quadruped robot swing leg in a complex environment, a spline optimization algorithm based on swing leg path planning was proposed. In this algorithm, the zero moment point (ZMP) stability criterion was used, the path points of the foot end trajectory of the robot swing leg were optimized on the basis of planning the COG trajectory of the robot, and the simulation model was established by using ADAMS for computer simulation. The results show that by using this algorithm, not only the safety of the quadruped robot to avoid obstacles can be ensured, but also the smooth walking under complex terrain conditions can be realized, the accuracy and robustness of the proposed algorithm are verified.