Abstract: This paper makes use of the principles and methods of semi-parametric analysis to construct a semiparametric model for house price indices by including the characteristics of house sale data. Further discussions axe given to the estimation methods of the semi-parametric model, Hausman hypothesis test and the use of bootstrapping to calculate the standardempirical analysis show that the semi-parametrichouse price indices.errors of the estimated parameters. Results of analysis is better than the OLS analysis in estimating house price indices.