Abstract: At present, there are two main ways of interpolating complex curves, which are fitted by the equidistant variation of the B-spline basis function parameters or the string approximation for velocity and acceleration control, but the fitting accuracy is not high. In order to improve the fitting precision, the local characteristics of the spline curve are considered comprehensively, and the error augmentation objective function is constructed. The particle swarm optimization algorithm is used to find the optimal solution for the uniform cubic B-spline curve interpolation trajectory. The results show that the algorithm can quickly converge to find the optimal trajectory, and effectively avoid the problems caused by local optimization.