Compared with conventional energy, new energy has the advantages of renewability, low pollution, and abundant resources, in which wind-power generation has been introduced into the micro-grid system on a large scale. However, due to the strong uncertainty of its generating capacity, how to improve the consumption level of new energy, improve the operation stability of power system and reduce the cost of micro-grid configuration has become an important and challenging problem. Aiming at the above problem, this paper presents a collaborative optimization method for energy storage capacity of wind-solar-battery complementary micro-gird system considering reliability cost, as well as the multiple time scale uncertainty of the wind light resources. This method adopts the autoregressive moving average model and Homer software simulation of discrete sequence of wind speed light irradiance, and builds out fan and PV generator output model; The expected value of power shortage is used as the evaluation index to establish the reliability cost function, and the cost model of micro-grid configuration is constructed. The improved second-order oscillation particle swarm optimization algorithm is used to calculate the optimal value of the wind-solar-battery complementary micro-grid system. The results show that the proposed method is feasible and economical, which can provide a reference for the configuration planning of the wind-solar-battery complementary micro-grid system.