The model-based approaches are difficult to be applied in practice due to the lack of formal models describing behaviors of systems. This paper proposes a method for generating accurate models of components in order to solve this issue in the context of the component-based systems. The key idea of this method is to generate minimal deterministic finite automata as the accurate models of the corresponding components. For this purpose, the proposed method first computes a set of traces as a regular language of a given component by executing all possible experiments over the alphabet of the component. This method then constructs a regular expression to represent this set. After that, a minimal deterministic finite automaton as an accurate model of the component is generated by applying the Thompson algorithm and some optimized activities. The generated models are useful for the existing model-based approaches, e.g., model checking and model-based testing in improving quality of component-based software. An implemented tool supporting the method and experimental results are also presented.