Simple algorithms for identification of inverse response models from step response are difficult to obtain because analytically the solution of a system of coupled nonlinear equations is required. In this article we propose a simple identification procedure for second order inverse response processes, based on the plant step response. The algorithm provides the model parameters in a sequential way, thus avoiding the solution of a nonlinear equation system. Moreover the algorithm is flexible because it can be suited to user requirements, thus modifying the algorithm performance. Finally error bounds on the identified parameters are provided which are useful if the model is used for control design purposes.