As an important parameter for the credit risk management, default probability is the basis for the calculation of expected default loss, bond pricing and credit portfolio management. In order to estimate default probability more accurately and effectively, this paper defines the incomplete information with the delayed filtrations to change the information structure and filtrate the noise so as to increase the accuracy of the estimation of default probability and replaces martingale with respect to the natural filtration produced by the Brownian motion with the delayed filtration produced by the doubly stochastic Poisson process to calculate the default probability so as to avoid the limitation that the movement of some financial assets is not a martingale. With this key of the credit model, default intensity can be estimated from market information. The method put forward in this paper can be applied for emerging markets without sufficient historical default data.