Partial shading condition makes the tracking speed and accuracy of maximum power point tracking difficult to guarantee. The Cuckoo Search Algorithm (CSA) and the Improved Perturbation and Observation (IP&O) method of adaptive variable step are studied and applied to the MPPT control of the PV system. The excellent global search capability of the CSA is used to converge near the global Maximum Power Point (MPP) quickly, then the excellent local search capability of adaptive variable step Perturbation and Observation (P&O) is used to converge to the MPP quickly and accurately. Finally, several lighting conditions are set up for simulation and P&O and Particle Swarm Optimization (PSO) algorithms are compared. Simulation results indicate that the proposed method has faster tracking speed and higher accuracy. This work is supported by National Natural Science Foundation of China (No. 51708194 and No. 51507014) and Key Project of Scientific Research for Hunan Education Department (No. 18A120).