Aiming at the output of photovoltaic arrays appears multipeak caused by the complexity of the environment in practical applications, the traditional control methods have difficulty in tracking the maximum power point of multipeak targets. This paper combines the optimized multipeak support vector machine function with the cat group algorithm with fast convergence characteristics, and proposes an improved cat group algorithm. The algorithm is used to solve the multipeak optimization problem in the case of partial and hence realize the maximum power tracking function of the solar power system. By constructing a mathematical model of photovoltaic cells, the multipeak output characteristics of photovoltaic (PV) cells under local shading conditions are analyzed. By analyzing the optimized structure of adaptive weights and mutation operators in IMCGA intelligent algorithm, this paper elaborates the implementation process of the IMCGA algorithm in maximum process point tracking of the shading photovoltaic system. Finally, the superiority of the proposed control algorithm is verified by simulation experiments.