This paper considered a design method of a reference governor which is used to improve control response by changing the reference input to the controller without changing the controller under the input restriction. In previous studies of reference governors, online reference governors using model predictive control have been proposed. However, it is necessary to solve a constrained optimization problem repeatedly to apply such online reference governors. To solve this problem, we employed the particle swarm optimization method which is an evolutionary computation method known for its fast solution search to design online reference governor. The effectiveness of the proposed method was confirmed by a numerical simulation.