This paper presents a method to establish functional inequalities via fuzzy decomposition on the state space, which generalizes earlier results dealing with exact partitions of the state space. Given a reversible Markov chain on a finite state space, we define its projection chain and restriction chains from classes of a fuzzy partition on the state space. The Poincaré, log-Sobolev and modified log-Sobolev inequalities associated with the original chain can be estimated from those of its projection chain and restriction chains which tend to have simpler structures and hence easier to work with. An application of this generalization is presented in which the fuzzy decomposition method applies but its exact partitioning counterpart does not.