Air pollution is one of main challenging issues nowadays that researchers have been trying to address. The emissions of vehicle engine exhausts are responsible for 50 percent of air pollution. Different types of emissions emit from vehicles including carbon monoxide, hydrocarbons, NOX, and so on. There is a tendency to develop strategies of engine control which work in a fast way. Accomplishing this task will result in a decrease in emissions which coupled with the fuel composition can bring about the best performance of the vehicle engine. Controlling the Air-Fuel Ratio (AFR) is necessary, because the AFR has an enormous impact on the effectiveness of the fuel and reduction of emissions. This paper is aimed at reviewing the recent studies on the prediction and control of the AFR, as a bulk of research works with different approaches, was conducted in this area. These approaches include both classical and modern methods, namely Artificial Neural Networks (ANN), Fuzzy Logic, and Neuro-Fuzzy Systems are described in this paper. The strength and the weakness of individual approaches will be discussed at length.