Water eutrophication, which refers to the enrichment of nutrients to an aquatic environment, is one of the most challenging problems in water protection. Although many researchers have made attempts to solve the eutrophication problem, there is one issue that needs to be further discussed, i.e., how to establish a fast, low-cost, and accurate eutrophication evaluation model? For addressing this issue, this paper proposes a data-driven eutrophication evaluation model based on the semi-supervised classification. Concretely, a case study in Three Gorges Reservoir of China is carried out to demonstrate the validity of the proposed model. Experimental results clearly show that the proposed model has the advantages of high computational efficiency, high accuracy, and great ability of exploiting low-cost factors to assist or even replace high-cost factors in realizing the eutrophication evaluation. Moreover, we find that three low-cost factors, including pH, dissolved oxygen, and ammonia-nitrogen, are effective in achieving a better eutrophication evaluation for Three Gorges Reservoir based on the proposed model.