Durability of a product is related to three major factors, the load, structure and material. The durability performance of an automotive product is, therefore, not only depended on the structure configuration, but also on the road load dynamic characteristics (profiles and frequency spectrum), and the material fatigue properties as well. Due to the dynamic nature of vehicle loads, one of the major technical challenges, to the durability design optimization of automotive products, is how to define a set of representative road loads, for fidelity and efficiency, based on the measured proving ground durability data of huge size.This paper presents a procedure of processing the proving ground road loads, for vehicle durability design and optimization of automotive products, based on the statistical characteristics evaluation and fatigue damage equivalency techniques. A practical method for constructing a set of representative road load data, with much shorter time duration, is introduced, which satisfy all statistical property fidelity requirements with respect to the original measured load data.The frequency spectrum and statistical characteristics of each road load event are first established from the measured proving ground data, by using the fast Fourier transform and power spectral density techniques. Based on the durability testing schedule and the fatigue damage level, the road load key life test events for lab tests and CAE are identified. For each selected road load event, a much shorter duration load data is then constructed, by iteratively searching and truncating a segment from the original measured load time history, based on the proposed technique. The resulted representative road load data, though is much shorter in time duration, maintains all statistical characteristics similar to that of the original measured dynamic proving ground data.