Abstract: Permutation flow shop scheduling problem is widely applied in process and discrete manufacturing enterprises. An improved Jaya algorithm is proposed to solve the PFSP with the minimum makespan. In the improved Jaya algorithm, four individual updating schemes based on the best and the worst individuals are proposed, and local search for individuals is carried out through four neighborhood structures. Diversity control strategy is applied to ensure the diversity of population. The improved Jaya algorithm is used to solve the Car, Rec and Taillard benchmark instances, and the experimental results validate the effectiveness of the proposed algorithm.