Objective To improve the efficiency and stability of spatial optimal time trajectory planning for food sorting robot arms.Methods A time-optimal trajectory planning method for food sorting robot arms based on an improved coati optimization algorithm (COA) was proposed. A UR5 six-degree-of-freedom robot arm model was established, and the 3-5-3 polynomial interpolation method was used to construct the motion trajectory of the robot arms. Drawing inspiration from the physical phenomenon of optical refraction, the COA was improved. The improved COA (ICOA) was then applied to optimize and solve the trajectory planning problem of robotic arms while satisfying constraints on angle, angular velocity, and angular acceleration, aiming to minimize joint motion time as much as possible.Results Simulation results showed that, compared with three other time-optimal trajectory planning methods, the proposed method reduced trajectory execution time by 14.58%, 21.20%, and 4.18%, respectively.Conclusion The proposed method effectively shortens the robot arms' operating time and significantly improves their working efficiency.