Abstract: To identify the frequency domain in fractional order systems in noisy environments, we propose an identification method that combines the particle swarm optimization (PSO) algorithm with the recursive instrumental variable (RIV) method. First, we extend the RIV method to identify the frequency domain of a fractional order system. Then, we combine the anti-interference performance of the RIV method with the global optimization capability of PSO. We apply the PSO algorithm to estimate the order and apply the RIV method to identify the multinomial coefficients. Experimental results and the identification example of an electric network show that the proposed method is applicable to commensurate, as well as general, fractional order systems.