For LCL-type grid-connected inverters, an adaptive control strategy based on RBF neural network is proposed. This control strategy can improve the traditional control strategy responding to the influence of steady state error caused by output power changing. When the external conditions change, the control strategy can modify the controller parameters constantly and get the optimal PI value. This control strategy can enhance the system stability and robustness. In order to solve the resonance of LCL filter, the capacitor-current-feedback active damping is adopted. Based on the amplitude and phase characteristics of the system, a design method of the capacitor-current-feedback active damping coefficient is presented. Simulations of LCL-type grid-connected inverters in grid-connected mode are conducted. Simulation results validate the correctness and feasibility of the proposed method.