For accurately predicting the correlation between the water level and output voltage for the point absorber wave energy converter (WEC), a nonlinear modeling study of the point absorber WEC by using AdaBoost-back propagation neural network (AdaBoost-BPNN) is reported. This paper tries to avoid the internal complicated mechanism of the WEC and presents a black-box identification model of the WEC. Simulation results have illustrated the applicability of the established AdaBoost-BPNN model in predicting the voltage characteristic under different water levels for the WEC. Furthermore, the comparisons between the AdaBoost-BPNN model and the BPNN model are provided which show a substantially better performance for the AdaBoost-BPNN model. Based on this model, performance predicting and controller design for maximum power extraction of the WEC can be developed.