In order to determine the state of lake eutrophication, starting from the attribute recognition theory, the paper established the attribute recognition model for the eutrophication assessment of lake, and combined with the weight coefficients determined by entropy method to evaluate the state of eutrophication of Dongchang lake. The result is that the lake district throughout the year belongs to eutrophication level, and the highest degree of eutrophication is July, followed by December, the lowest degre of that is April. This result is consistent with the result evaluated by set pair analysis and fuzzy recognition method. The evaluation result of the model is reasonable. The model is simple,easy to be understand and can be used as a feasible way for the eutrophication assessment of lake.