Structural Identification (SI) is basically an inverse approach for predicting system dynamic parameters from input output observations. This work presents a novel nonlinear lap joint parameter identification method based on instantaneous power flow balance approach, which includes structural nonlinear parameters for realistic modelling. In the present approach, substructure instantaneous power flow balance whereby algebraic sum of input power, dissipated power, transmitted power and time rate of kinetic and strain energy is equated to zero, is used as an objective criterion to formulate an inverse problem for nonlinear joint parameter identification. The correct values of the nonlinear coefficients are estimated by minimizing the objective function using Particle Swarm Optimization algorithm. The method is first explained using simulated experiments on a lap joint assembly formulated using a generic element. After the feasibility study the method was applied to experiments involving a steel beam assembly connected by a single bolted lap joint with various tightening torques. Both the simulated and experimental structure applications have shown that the proposed method is effective for nonlinear parameter identification of joints.