In order to solve the shortcomings of the improved distributed algorithm such as low efficiency of late optimization search, an improved distributed algorithm for optimization of microstrip bandpass filters for ultra-wide band wireless communication is proposed. The algorithm is applied to the optimized design of the microstrip ultra-wide band filter, and the corresponding optimization model is established, which is modeled and simulated in the high frequency simulation software. The experimental results show that the microstrip ultra-wide band filter designed in the research is done in a microprocessor at 2 GHz, and the approximate computation time takes only 1.2 min. The improved distribution needs to go through 12 iterations to obtain the optimal solution. In terms of passband and stopband attenuation, the improved distributed algorithm has less fluctuation than the other two optimized algorithms. The stopband attenuation is 43.58 dB, which is 15.37 dB higher than that of the genetic algorithm. The effectiveness of the method is verified in the research.