This paper describes techniques that predict and analyze dynamic response of vehicles traversing random rough surfaces. Road irregularities are statistically classified by frequency and amplitude distribution. This classification determines the nature of random inputs to mathematical vehicle models and allows computer prediction of dynamic response of a simulated vehicle. Once inputs and models are defined, parametric analysis with output criteria specified statistically can be performed. This allows prediction of vehicle riding quality and evaluation of design concepts. Statistical analysis of accelerometer measurements on actual vehicles permits verification of the design process and meaningful comparison between vehicles.