Abstract: Most of the current reliability analysis methods based on surrogate models combined with Monte Carlo Simulation (MCS) depend on the Kriging model. The reliability analysis methods based on other surrogate models are still relatively rare. One new structural reliability analysis method is proposed with active learning, MCS, and ensemble Support Vector Regression (SVR) models with different kernel functions. The method proposed utilizes the Inter Quartile Range (IQR) to estimate the local prediction uncertainty of the SVR ensemble model, based on which an active learning function is constructed to control the adaptive updating of the SVR ensemble model until convergence. The effectiveness and accuracy of the method proposed were verified against three examples. The analysis results show that the method has good applicability and high efficiency for multi-failure- domain, nonlinear, and high-dimensional problems. Although the ensemble model increases the modeling cost compared with the single model, the overall computational efficiency and accuracy are better than those of the conventional method based on Kriging model for high-dimensional problems.