To address the problem of modelling under the condition of random parameters using Probit quantile regression, this paper proposed Bayesian Probit quantile regression model based on the Metropolis-Hastings algorithm. According to the Probit quantile structure, the M-H algorithm was ulitized to simulate the posterior marginal distribution by choosing the prior distribution of the parameters. The model posterior distribution in the different quantile points was obtained by using Monte Carlo simulation, and at the same time, Probit quantile regression, smoothed Probit quantiel regression and Bayesian probit quantile regression method were used to estimate the parameters of the model and to compare the differences from the estimation of parameters, respectively. The research shows that Bayesina Probit quantile regression comprehensively describes the influencing factors of discrete variables, and the estimation of parameters is more accurate and efficient.