由于传统的CART决策树模型存在运行时间较长和预测精度不够等问题。改进CART决策树利用Fayyad边界点判定定理，减少挑选属性最优阈值所用的计算时间，减少整体的运行时间。由于影响水华生成的因子较多，再利用统计学中的相关系数选出与水华发生的相关性较大的影响因子，提前一步筛选条件属性，进一步缩短运行时长，并且能够保证整体的预测精度。将这种改进了的CART算法用于生成湖体水华预警模型。最后实验结果表明，改进后的水华预警模型能减少运行时长并很好的保证预测的正确率。

As the operation of traditional CAＲT decision tree model is time-consuming，we choose the Fayyad Boundary Decision Theorem optimization attributes to reduce the time for operation． Because of the factors that influence the formation of bloom，this paper uses the correlation coefficient in statistics to select the influencing factors that have a greater correlation with the occurrence of bloom，one step is to filter the condition attributes one by one，to further shorten the running time，and to ensure the overall prediction accuracy． Finally，the experimental results show that the improved bloom warning model can reduce the running time and ensure the correct rate of prediction．