以正交设计方法为基础，根据CERES-Maize模型品种参数的实际意义和气象资料、品种特性，确定了各遗传参数的取值范围，利用正交表设计了不同参数组合的正交试验，以模拟产量、发育期的准确率为参考指标进行方差分析，从单因素统计结果中选择平均值最大的水平为最优方案，确定出每个样本的遗传参数，根据不同品种熟性分为晚熟品种、中熟品种、早熟品种3类，选择各类遗传参数中出现频率较多的参数组合成能代表不同熟性品种的遗传参数。利用实测资料进行验证，CERES-Maize对产量、开花期和发育期的模拟效果较好，大多数年份的误差在±10％以内，个别年份在±10%～20％，模拟结果比较合理。

Based on the orthogonal design, the methods of establishing maize variety parameters was set up in the CERES-maize simulation model. First of all, the primary scale of parameters was made based on the meaning of parameters, meteorological data and variety characteristic. Then the orthogonal experiment was made use orthogonal layout. The variance was analyzed according to the accuracy of yield, physiological maturity. The optimizing project was selected from the single factor statistic. We got the parameters of every sample. The varieties was divided into long season, medium season and short season. At last, the parameters of different seasonal varieties were made sure according to the frequency. The validation showed that the simulation of CERES-Maize was good to the yield, flowing data and physiological maturity, the error of most samples were less than ±10％. Therefore, the selection of parameters was reasonable.

国家科技攻关计划项目(2001-BA611B-04-06-02)、中国气象局气候变化专项(CCSF2007-27)