This study investigates the sampled-data reliable control for a class of Takagi-Sugeno (T-S) fuzzy semi-Markovian jump system (FSMJS). The system under consideration involves semi-Markov stochastic process that is described by the property of cumulative distribution function. To relax the difficulty of finding upper bounds in sample-and-hold behaviour of the FSMJS, a novel method named mismatched membership function is presented. Sufficient conditions that can guarantee the T–S FSMJS to be stochastically stable are given. More precisely, the reliable sampled-data controller is designed in terms of linear matrix inequalities. Finally, the validity of the presented approach is demonstrated by a controller design for a single-link robot arm model.