Formulating accurate dynamic load models is critical for power system analysis, control, and planning. In this paper, a generic dynamic load model is proposed. The dynamic power response of the load is directly approximated as the superposition of various mathematical functions that produce a dynamic response. Basic physical principles of the dynamic process are reflected in the mathematical functions utilized in the proposed model. First, different stages of the dynamic process are detected based on the continuity of derivatives of the measurement. Second, a complete set of mathematical functions that produce the dynamic responses in electric devices are formulated. Third, a parsimonious set of mathematical functions is selected at each stage by performing feature selection using nonlinear principal component analysis. The proposed model is further formulated based on the parsimonious set of mathematical functions at each stage. Finally, the parameters of the proposed model corresponding to different system events are solved. Based on the formulated model, its possible application in event detection is further analyzed. The proposed model is easy to implement given limited data measurement. Various tests on different system event data are performed to validate the effectiveness of the proposed model. The results show that the proposed model has excellent accuracy and robustness for different system events.