As constrained optimization problem can not be solved by existing robust optimization over time , this paper proposes a new algorithm for robust optimization over time for solving constrained optimization problems based on swarm intelligence. Firstly, a penalty function with constraints is constructed and the constrained optimization problems are processed into the unconstrained optimization problems. Then, the algorithm of robust optimization over time for constrained is designed, which the piecewise function is used as the fitness evaluation function of the particle, and the particle is selected by the competition rule. Finally, The proposed algorithm is evaluated and compared under different parameter settings for optimization of the performance of a carbon fiber precursor to demonstrate its effectiveness. Furthermore, the influence of AR model on the performance of the algorithm is analyzed. It is pointed out that the improvement of the prediction model is an important way to improve the performance of the algorithm.