Abstarct: Roughness is a major index in order to achieve ride comfort and road safety. For a driver, driving on rough roads means lack of comfort, the slowdown and possible damage to the vehicle. One of the reasons for the creation of roughness are cracks and potholes. RSP is one of the profilometer which gives the longitudinal profile of roadway. This simple profile can be filtered with the various methods to achieve various goals. The method which is used in this study is wavelet analysis. This research at the first, studied the principles and concepts of roughness introduced different equipments in order to measure roughness; then introduced filteration methods and how to filter roughness data by one of these methods (wavelet analysis). In conventional methods, noises are removing from the original signal and no analysis has been done on it. Part of these noises are caused by cracks in the road surface; therefore, in this study these noises have been used to detect a variety of cracks. In the following set of new parameters will be presented so that they can make a logical relationship with cracks and potholes in each part of pavement. These parameters are as follows: the original signal, the approximation component and the details component, the slope characteristic of details component, the hight characteristic of details component, the average hight characteristic of details component for 6 different interval length (2, 3, 4, 6, 8 and 10 units of length). At the end, the resuts showed that the model which is presented in this study can recognize the transverse cracks, block cracks and alligator cracks; but this model dosen't have an ability to detect pothols and longitudinal cracks.