Recommendations of crop varieties are usually done on the basis of grain yield. However, there are other important traits related to quality and agronomy that need to be considered. This study has clearly shown that grain yield is often not related to some important traits of the crop. Under such circumstances, selection based on yield can lead to the loss of these important characters. Recent research has also indicated the need for a multi-trait approach to recommending crop varieties. However, such approaches, as suggested in the literature, are subjective and do not have a sufficient statistical basis. This study proposes a methodology for varietal recommendations by taking account of all important traits. It is a multivariate approach considered to be an improvement on the univariate method previously proposed. Data from rice varieties of 3, 3½ and 4½ month maturity groups, cultivated across diverse environments, over two seasons in Sri Lanka were used to illustrate the proposed methodology. The results suggest that the new method will be appropriate for taking in to account all important traits along with yield.