Evaluation of road pavement is an important part of maintaining its quality. In a traditional way, human checks the asphalt by walking along the road. This traditional method it is less efficient because it requires substantial costs, takes a long time, exposed to safety issues such as the high intensity of vehicle passing by, the subjective factor, and the fatigue factor. With digital image processing technique, pavement evaluation will be safer for surveyor. Pavement’s cracks are captured on a picture and then processed with some algorithm. This paper explains the principle of Discrete Wavelet Transform (DWT). The image is filtered by the low-pass filter and a high-pass filter and decomposed into four subbands to perform feature extraction to extract unique features of the picture. These extracted features are then applied to classification method using linear discriminant analysis (LDA), and the accuracy is about 92.8%