本文对作物品种区域试验中各试验场点多年非平衡数据，提出了相应的数学模型和统计方法．以渐江省晚粳（糯）稻品种区域试验资料为例作了分析，并对区试中各试验场点评价指标作了探讨．分析结果表明，浙江省晚粳（糯）稻品种1989～1995年区域试验中，以渐江省诸暨稻麦原种场、舟山试验点综合评价效果最好，其次是富阳、龙泉、台州、宁波、衢州试验点． In this paper, the multi-year unbalanced data of various test sites in the regional trials of crop varieties were put forward, and corresponding mathematical models and statistical methods were put forward. Taking the regional experiment data of late japonica (waxy) rice varieties in Qianjiang as an example, the evaluation indexes of each test field in the district test were discussed. The results showed that the comprehensive evaluation was the best in Zhejiang province Zhuji rice grain and wheat wheat field and Zhoushan experimental site in the experiment of late japonica (waxy) rice varieties in 1989 ~ 1995, followed by Fuyang, Longquan, Taizhou, Ningbo Quzhou test point.