We investigate the fractional-order systems which are perturbed by stochastic input to achieve stabilization via sliding mode control (SMC) approach. It is assumed that the system states are unknown and there is uncertainty and time-delay in the system. We utilize the diffusive representation of the stochastic fractional-order dynamics to transform the system into an integer-order system perturbed by Brownian motion. Provided that some linear matrix inequalities (LMIs) are feasible, it is proven that the estimation error system is stochastically stabilized and the overall closed-loop system is stable in probability. A numerical simulation shows the effectiveness of the results.