Evaluation of road pavement is an important task to maintain its quality. In a traditional way, officer checks the road surface by surveying along the road. This traditional method is less efficient because it requires extensive costs, takes a long time, exposes to safety issues, and less accurate due to human subjective factor and fatigue. The objective of this research is to develop a features extraction method based on wavelets to detect crack and non-crack road surface. The method involves road surface acquisition, pre-processing, features extraction using wavelets and classification task using linear discriminate analysis. The developed method was implemented on 56 images and produced 92.8% of accuracy detection of crack and non-crack