The classification of contaminated sites based on human health as well as ambient biological and ecological risk evaluation has become an important tool to deal with increasing needs for risk management and remediation for soil and ground water pollution of contaminated industrial sites. We established 40 indicators of three tiers involving pollutants’ characteristics, immigration pathway and recipients, and developed an integrated fuzzy evaluation model for classification of the contaminated industrial sites by applying triangular fuzzy numbers to describing evaluation indicators. We also examined the criteria for determining risk grading and risk importance and the values of risk grading, and the risk importance could be obtained based on site investigation information. Through multi-layer risk rector analysis, a comprehensive conclusion could be drawn as a reference for prioritizing site rehabilitation action. The model built with MATABLE program was applicable in evaluating the site of a pesticide factory, and the results showed that the site was at high risk level (V8) and timely rehabilitation would be necessary.