This paper investigates the efficient model predictive control(EMPC) problem of a class of interval type-2 T-S fuzzy systems with deception attacks. Considering the influence of deception attacks during the data transmission from the sensor node to the controller node, an observer-based output feedback controller in the framework of EMPC is designed. Firstly, aiming at the energy bounded deception attacks, an auxiliary optimization problem is established. Then, by solving this problem offline, the terminal constraint set and the gains of the corresponding fixed feedback controller and observer that guarantee the system $H_2$ security are derived. Secondly, perturbations are employed to provide more control freedom, and the invariant set of the augmented system can be designed off-line. In this sense, by maximizing the projection of such an invariant set in the original state space, the initial feasible region is readily obtained. Thirdly, an online optimization problem is put forward to find the desired control law to guarantee system security under deception attacks. Finally, the proof is provided to show the theoretical feasibility of the proposed EMPC algorithm and the security of the closed-loop system. In addition, the effectiveness of the proposed algorithm is verified by a simulation example.