A fuzzy pattern recognition model based on projection pursuit for lake ecosystem health evaluation is proposed in this paper. There are four procedures in this method: (1)screening a set of indicators that can reflect the health status of lake ecosystem, (2)building the evaluation criteria, (3)determining the index weight using the projection pursuit method, and(4)evaluating the lake health status through adoption of the fuzzy pattern recognition model. The model was applied to the health assessment of Hongze Lake. The results suggest that Hongze Lake was in a slightly polluted state. The high concentrations of TN and TP were the main factors that affected the health status of Hongze Lake. The evaluation results of the model corresponded with the lake’s actual health status. Therefore, the model can be widely applied to the evaluation of other ecosystems.