In view of the critical flutter of wind turbine blades at flutter speed, a circular secant modified particle swarm optimization (CSM-PSO) algorithm was designed by combining the geometric circular secant with the traditional PSO algorithm.The proposed CSM-PSO was applied to the parameter identification of a blade critical flutter system.In the method, the moving circumscribed distance was used to adjust the global learning factor and the local learning factor to improve the dynamic balance between the global search and the local search for the optimization identification.It is found that in this way, the algorithm can avoid falling into the local optimization and improve the overall optimization performance and computational efficiency.In simulation experiments, the results of the method were compared with the results of existing modified PSOs, such as modified particle swarm optimization(MPSO), linearly decreasing inertia weight based particle swarm optimization(LDIW-PSO) and immune particle swarm optimization based on dynamically changing learning factors(IPSODCLF).The results show that the method has significant improvement in identification accuracy, computation time and robustness.