This paper investigates robust fault estimation and fault tolerant control problems for stochastic Lipschitz nonlinear systems subject to Brownian motions, unexpected faults and unknown inputs. Augmented system approach and unknown input observer are integrated to produce robust estimates of the means of the faults and the full system states simultaneously. Based on the well-designed fault estimation scheme, a robust fault tolerant control strategy is proposed to compensate faults, stabilize the closed loop system, and eliminate the effects of unknown inputs. Sufficient conditions are presented in terms of linear matrix inequalities for the overall closed-loop system composed of both states and error dynamics to guarantee the stability and robustness of the system. Furthermore, the systematic design procedure for the robust fault estimation and fault tolerant control scheme is addressed. Finally, simulation on a single joint robotic model is illustrated to validate the suggested methodologies.