The solar array whose quality has great influence on the work efficiency of the spacecraft is the most important power supply system and the honeycomb sandwich panel which has high specific strength and stiffness is the perfect material for solar array.The finite element model(FEM) of the honeycomb sandwich panel includes idealized hypothesis and simplification,which results in the difference between the calculating data and the test data.In this paper the honeycomb sandwich panel is equivalent to a structure which can be calculated by NASTRAN based on the equivalence theory and then the improved Genetic-Partical Swarm Optimization ( GA-PSO) algorithm is used to update the FEM,the updated precision and efficiency of the GA,PSO and GA-PSO are compared and the results show that these random search algorithms can update high order modes even use the incomplete low order modes.The improved GA-PSO algorithm can use GA to efficiently search for the global optimal solution at the early stage,and use PSO to intensively search for the local optimal solution at the later stage,it has higher searching efficiency and precision,the updated model has good apprcaching and prediction ability and can be taken as the fiducial model for the following calculation.