This paper applies an multi-factor smooth transition regressive model to capture the regime-switching behavior of the relation between credit default swap(hereafter CDS) spread and macroeconomic variables. We have divided the type of time series in the sense of previous studies(level, transitory component, difference). The results are as follows: First, in preliminary analysis as linearity test using the Taylor expansion, we find the non-linear relationship between CDS spreads and explanatory variables which is consistent with the smooth transition regression used in this study. Second, CDS spread is responded the most by the shock of exchange rate of among all the macroeconomic variables, followed by the shock of risk-free rate. Third, according to the market regime, macroeconomic factors have different effects on CDS spread. Fourth, In prediction section, the coefficients of lagged exchange rate and risk-free rate are statistically significant. The fact that lagged exchange rate and risk free rate are significantly related to CDS spreads may be interpreted as evidence of inefficiency in CDS markets. Finally, STR`s adj - R2S are larger than linear model`s adj - R2S on all analysis. These results indicate that smooth transition regressive models have a better fit.