This paper analyzes the connection between the money market and the capital market in a micro level sense，investigates the connection approaches and their pattern structures，and mines deeply the dynamics of interactions between the two markets through the integration of support vector machines ( SVM) into copula functions． The modified sample-weighted SVM is employed to estimate the marginal probability distribution functions of the money market price，SHIBOＲ，and the stock price index return of the capital market in terms of the nonlinearity and instability of financial variable time series，and the optimized copula function is then used to analyze the joint probability distributions of the two financial variables and their change scenarios under various circumstances，with the dependence structures between the two markets and the dynamics of the nonlinear connection obtained． The empirical analysis shows that the joint distribution probability of 1Y-SHIBOR and the stock price index return exhibit different structure characteristics under different directions and magnitudes when changing the two variables above，there exists asymmetries，and the stress test analysis gives similar results．