Mortgage insurance (MI) is a contract under which an insurance company has an upper limit to compensate the default losses of mortgage banks or investors. Previous studies evaluate the MI premium under the Black-Scholes framework. However, the returns of the HPI exhibit housing price cycles, asymmetry and volatility clustering. In this paper, we utilize the Markov regime-switching framework which is more suitable than Black-Scholes model to address these characteristics of return. Finally, based on the sensitivity analysis, the housing price cycles of the HPI return is an important factor that influences MI premiums.