Photovoltaic grid-connected inverter system has the advantages of simple topology and low cost. Since the output power of photovoltaic devices is a nonlinear function of the external environment load, to maximize the performance of photovoltaic devices, Maximum Power Point Tracking (MPPT) control should adjust the duty cycle disturbance based on the work of photovoltaic devices. Although MPPT control with using conventional perturbation and observation method has the advantages of simple in structure and easy to implement the hardware circuit, but the control is imprecise and general poor. Therefore, aiming at the potential inadequacies of using the traditional MPPT control algorithm in a single-stage photovoltaic inverter system, in this paper, the traditional MPPT algorithm is proposed based on a new control algorithm, the artificial fish swarm algorithm is applied to MPPT control of a single-stage photovoltaic grid-connected system. Parameter settings of AFSA are changed through simulation to seek the optimal power value, and this control method is compared with the traditional MPPT control method P&O, simulation and experiment have demonstrated the reliability and effectiveness of the algorithm.