In 1987, Eitelberg introduced a useful method by adjusting the input weighting factors to recover system response when a direct tuning of the PI(D)'s parameters was prohibited from adjustment. Because of minimum interruption to an original control system, the Eitelberg's method has been extended to various applications. From previous applications, it was found that the P-element input weighting factor distinctly effects on the system responses. In this paper, the input weighting factor optimization for P-element of PI controllers via the particle swarm optimization (PSO) method is presented. The proposed optimization method is tested against two benchmark-controlled systems. As the simulation results, it was found that the P-element input weighting factor can be appropriated optimized by the PSO-based method and the satisfactory responses can be recovered when the parameter variation occurred.