When the traditional whale-conductance increment method is applied to photovoltaic system under complex condition, the P-U characteristic curve has the problems of large vibration amplitude and long convergence time, which restricts the energy conversion efficiency and working stability of the system. In view of the above problems, an improved compound MPPT algorithm combining whale algorithm and conductance increment method is proposed. In this algorithm, nonlinear convergence factor containing adaptive weight is introduced to replace the linear convergence factor of the original whale algorithm, so as to ensure that the algorithm can leap out of local optimal and enhance global search ability. Meanwhile, the algorithm adds a screening perturbation mechanism to improve the optimization speed while accurately optimizing. The simulation results show that the proposed algorithm can converge faster than the traditional whale-conductance increment algorithm. The optimization speed is significantly faster, and the output curve is smoother.