Abstract: In the process of oil gas exploration, it is difficult to make a quantitative evaluation for hydrocarbon potential of a geological target, and there are too many factors having different influence such as trap, source, deposition, preservation, temporal and spatial matching, and lots of other smaller factors are included. Multi-level fuzzy comprehensive evaluation is successfully applied to solve this problem. The procedure of ranking 4 secondary tectonic units in Dongying Depression is given as an example. The outcome of computation indicates that the northern steep slope and the central uplift are the best for hydrocarbon accumulation, and that the southern gentle slope is the second and the subsag the worst, which tallies with the latest exploration findings in this area.