为更好选择和利用优势亲本，提高水稻新品种选育效率，以来自福建、东北地区和宁夏的60份水稻种质为材料，通过变异系数、遗传多样性指数、聚类分析、主成分分析、逐步回归分析等方法对水稻18个主要表型性状的多样性进行评价。结果表明，18个表型性状的变异系数在1.10%~63.40%之间，其中黄粒米的变异系数最大（63.40%），整精米率、垩白粒率、垩白度、不完善粒等稻米外观品质变异系数次之，均在34.00%以上，出糙率和生育期的变异系数较小，分别是1.10%和2.94%；18个性状的表型多样性指数为1.74~2.10，其中生育期的多样性指数最低，小区产量的多样性指数最高；按照来源对60份稻种资源表型多样性指数进行分析，各地区稻种资源的表型多样性指数变异较大，其中福建的变异幅度最大（1.06~2.10）、东北次之（1.50~2.09）、宁夏变异范围最小（1.51~2.01）。参试水稻资源在遗传距离为1.03时分4类，明确了各类资源的特征特性；基于主成分分析的综合D值，对种质资源进行排序，筛选的综合得分最高的前10名种质；同时利用主成分分析的D值与逐步回归分析，构建了水稻资源评价方程，该方程的构建为水稻资源综合评价、筛选提供了依据。

In order to better select and utilize the dominant parents and improve the breeding efficiency of new rice varieties, 60 rice germplasms from Fujian, Northeast China and Ningxia were used as materials, the phenotypic diversities of 18 main traits of rice resources were evaluated through the coefficient of variation, genetic diversity index, cluster analysis, principal component analysis, stepwise regression analysis and other methods. The results showed that the coefficient of variation of 18 phenotypic traits was between 1.10% and 63.40%, among which the coefficient of variation of yellow-grained rice was the largest（63.40%）, the coefficients of variation of rice appearance quality such as the rate of whole milled rice, chalky grain rate, chalkiness degree and imperfect grain were relatively large, all above 34.00%, while the coefficients of variation of roughness rate and growth period were relatively small, 1.10% and 2.94% respectively. The diversity index of the 18 phenotypic traits was 1.74-2.10, among which the genetic diversity index of the growth period was the lowest, and the diversity index of the plot yield was the highest. The phenotypic diversity index of 60 rice resources was analyzed according to the source. The phenotypic diversity index of rice resources in various regions varied greatly, among which Fujian had the largest variation range （1.06-2.10）, followed by Northeast China （1.50-2.09）, and Ningxia had the smallest variation range （1.51-2.01）. The tested rice resources were clustered into 4 categories with the genetic distance of 1.03, and the characteristics of each type of resources were clarified. Based on the comprehensive D value of principal component analysis, the germplasm resources were sorted, and the top 10 germplasms with the highest comprehensive score were screened. At the same time, using the D value of principal component analysis and stepwise regression analysis, an evaluation equation of rice resources was constructed, which provides a basis for comprehensive evaluation and screening of rice resources.