This paper develops a new credit risk model for small and medium-sized enterprises (SMEs) based on the DSW model of stochastic default intensity and the dynamics of underlying time-varying covariates. In particular, our model incorporates the default probability with the probability of initial public offerings (IPOs) in the framework of a censored stopping-time model. Default stopping time comprises of such variables as total borrowings/total assets in the stability, ROA in the profitability, account payable/sales in the activity, and financial expenses/total cost in the etc. As for the IPO stopping time, the natural log of total assets in the stability, net income/net sales in the profitability, and cash flows from operating activities/ net sales in the etc. are significant for the IPO stopping time. It is found that our model based on DSW model outperforms the multi-period logit model consistently and robustiously according to various prediction horizons and lag orders of VAR for macro-variates because the continuous stopping-time framework emphasizing on the stochastic default intensity accurately calculates the default probability superior to the discrete-time model equivalently computing the survival and default probability. In addition, it captures countercyclically monthly-frequency movement of capital requirements in compliance with the new Basel Accord. The implication is that our model as the early warning system may help the financial supervisory authority to predict the expected loss due to the sudden collapse of economic fundamentals such as rapid transfer of debtors' credit risks.