All optimization methods used in optical thin film design can seek only one object. After the analysis of the theory of optical thin film optimization design, it is proposed that multiobjective optimization with mixed discrete variable is the essential mathematical model and can reflect the physical essence of optical thin film design. The optimization algorithm for single objective problem which has been widely used in the fields of optical thin film design is only a simplification. Based on this new consideration and the status of multiobjective optimization research, an immune response-based multiobjective optimization algorithm is adopted to design optical thin film. This algorithm can implement parallel treatment, and is a self-adaptive random algorithm with overall search capability in principle. The algorithm is used in optical thin film design, and some examples are presented. According to the results of experiments, the idea of applying multiobjective optimization approach to designing optical thin film can be realized in theory and have a bright future.