The stochastic volatility and self--exciting jump process are incorporated into a short-rate model. In the model, the self-exciting jump process will modeled by a Hawkes process, which captures the jump cluster. The expansion of the differential operator is applied to compute the closed-form moment function, and further develop the general moment method to estimate the parameters in the model and make statistical inference. The empirical results provide that there is no enough evidence to support the goodness of fit test. But the model with Hawkes process is significance in statistics, and  could strongly capture the jump clustering. Finally, the filtered values are estimated for the stochastic volatility, jump size, jump probability and intensity of jump by using filtering method. It is worth mention that the filtered jump probability is a plausible indicator to measure financial market stress.