Vehicle identification systems rely on logo recognition to identify vehicles (VLRS). Convolutional Neural Networks are used to automatically learn characteristics for car logo recognition (CNNs). However, CNN struggles with rotated or noisy pictures. CNN's Random Forest classification technique is used to create an image recognition system. Random forest decision tree ensemble and train. This work's primary contribution is a multiclass logo using convolution mapping in nonlinear space and random forest ensemble learning. In the experiment, 400 pictures with 10 classes were analyzed to increase accuracy by about 20%.