Summary: In structural reliability analysis, robust and efficient sampling methods that address low failure probabilities are vital challenges. In this paper, a novel dynamical adaptive enhanced simulation method coupled with support vector regression (SVR) is proposed. Firstly, a more general and efficient approximation formula is proposed as an improved scheme. Furthermore, a dynamical adaptive simulation strategy for Monte Carlo simulation and an active training methodology basis SVR are developed. The dynamical active region for improving the efficiency and robustness of structural reliability analysis is applied for training the SVR models which are utilized for accurate estimating the failure probability by simulation methods. Analytical methods and crude Monte Carlo simulation are used for comparison, validation and discussion with the proposed hybrid simulation method using four numerical examples and four engineering problems. Through coupling SVR with dynamical active region, an accurate failure probability prediction with robust and low-computational cost is achieved. The proposed adaptive strategy applied in hybrid enhanced simulation approaches provided the accurate results with low-computational burden and these hybrid methods are robust than the analytical approaches. The proposed methods have shown strong capability for application in engineering problems with complex nonlinear performance functions.