In this paper, a novel method, called the fuzzy-based approach, for mesh simplification is presented. We first propose the fuzzy position uncertainty function and the fuzzy curvature uncertainty function for, respectively, measuring the variation of the surface position and the variation of the surface curvature while meshes are simplified. We then utilize the TSK fuzzy inference model, which integrates and balances the fuzzy position uncertainty and the fuzzy curvature uncertainty, to determine the cost as a criterion for removing a portion of a mesh in the simplification process. Experimental results show that our approach can produce good approximations that preserve the features of a model.