Detection and recognition vehicles are important thing in systems and biling policy. in order to recognition the vehicle, it needs some features. this research using side view of vehicle image to recognize the vehicle from is wheels. the purpose of research is to built recognition model of wheel using circle hought transfrom (CHT), 2DPCA and artificial neural network (ANN). The CHT used to detect the circle of wheel in the image,2DPCA for feature reduction and ANN to recognize the wheels. the result of the research is the accuracy level of succes and the error for wheel recognition. the error consist of two pars including the miss and the false alarm rate. in this research, the highest accuracy of success is 94,4% with 100 neuron in the hidden layer, its also being tested data with noise and fake circle.