Air-fuel ratio control is a crucial problem for engine control since it is one of the most important issues related to pollution reduction. The main difficulty in air-fuel ratio control is the time-varying delay. We propose a new model that includes the delay. This model is identified using real dataset from an engine test bench. The time-varying delay is made constant by using a change of domain. The nonlinearities of the model are handled with the Takagi-Sugeno representation and a linear controller is designed using the Lyapunov direct method. Simulation results highlight the efficiency of the proposed approach compared to classic maps-based controllers.