Abstract: Aiming at the problem of capacity optimization configuration of wind/solar/hydrogen/storage grid-connected microgrid, a capacity optimization model with economy, energy utilization and power supply reliability as objective functions is constructed. Aiming at the problems of slow speed and low precision of traditional algorithm, an improved honey badger algorithm is proposed. In the initial population stage, the Bernoulli chaotic map is used to make the population have ergodic uniformity. In the mining stage and the honey-seeking stage, the piecewise nonlinear decreasing parameter is introduced to find the case where there are better individuals near the optimal individual. The horizontal crossover strategy is used to generate new solutions to enhance the search ability of the algorithm. Four test functions are used to compare the improved honey badger algorithm, honey badger algorithm, particle swarm optimization algorithm and butterfly optimization algorithm. The results of the example show that the improved algorithm reduces the cost by 7.44%, the energy waste rate by 21.79%, and the redundancy by 3.89%.