This paper presents a two-step algorithm for estimating dynamic games under the assumption that behavior is consistent with Markov Perfect Equilibrium. The first step involves estimating policy functions and the law of motion for state variables, while the second step estimates the remaining structural parameters using simulated minimum distance estimation. The algorithm is applicable to a broad class of models, including input-output models with both discrete and continuous controls. The authors test the algorithm on dynamic discrete choice models and dynamic oligopoly models, demonstrating its effectiveness even with relatively small data sets. The paper also discusses the computational efficiency of the algorithm and its ability to handle models that are not point identified.This paper presents a two-step algorithm for estimating dynamic games under the assumption that behavior is consistent with Markov Perfect Equilibrium. The first step involves estimating policy functions and the law of motion for state variables, while the second step estimates the remaining structural parameters using simulated minimum distance estimation. The algorithm is applicable to a broad class of models, including input-output models with both discrete and continuous controls. The authors test the algorithm on dynamic discrete choice models and dynamic oligopoly models, demonstrating its effectiveness even with relatively small data sets. The paper also discusses the computational efficiency of the algorithm and its ability to handle models that are not point identified.