The paper presents an application of the inverse analysis to the identification of two models: a phase transformation model and a rheological model. The optimization algorithm for the inverse analysis was tested for various techniques of searching for the minimum: derivative-free and gradient methods, as well as genetic algorithms. Simulation results were validated for microalloyed niobium steel. An optimization strategy, which is adequate for the inverse analysis, is suggested.