A pseudo-spectral control algorithm based on adaptive Gauss collocation point reconstruction is proposed to efficiently solve the optimal dynamics control problem of industrial robots. A mathematical model for the kinematic relationship and dynamic optimization control of industrial robots has been established. On the basis of deriving the Legendre–Gauss collocation formula, a two-stage adaptive Gauss collocation strategy for industrial robot dynamics control variables was designed to improve the dynamics optimization control effect of industrial robot by improving the solution efficiency of constrained optimization problems. The results show that compared with the control variable parameterization method and the traditional Gaussian pseudo-spectral method, the proposed dynamic optimal control method based on an adaptive Gaussian point reconstruction algorithm can effectively improve the solving time and efficiency of constrained optimization problems, thereby further enhancing the dynamic optimization control and trajectory tracking effect of industrial robots.