This paper studies fuzzy filtering design for a class of nonlinear discrete-time uncertain stochastic systems. First, the Takagi and Sugeno fuzzy model is proposed to approximate a nonlinear discrete-time uncertain stochastic system. Next, based on the fuzzy model, the fuzzy estimation for nonlinear discrete-time uncertain stochastic systems is studied. Using a suboptimal approach, the fuzzy estimation problem for nonlinear discrete-time uncertain stochastic systems is characterized in terms of an eigenvalue problem (EVP) by minimizing the upper bound on the variance of the estimation error. The EVP can be efficiently solved using convex optimization techniques.