This paper aims to evaluate and classify some upland rice varieties based on their morphological characteristics. The data used in the study was obtained from National Cereal Research Institute (NCRI), Badeggi -Nigeria. The study consists of 25 upland rice cultivars, which were extracted from 100 rice varieties ergo-listed in the institute. The experiment was conducted at Edozhigi in north- central Nigeria. The experiment was laid out in a randomized complete block design with three replications and sixteen agro-morphological characters measured on each of the experimental plots. Preliminary analysis of variance showed that the cultivars were significantly different (p< 0.05) in their plant height and leaf length and highly significant (p<0.01) for first and second iron toxicity score, panicle/m2, panicle length, panicle weight, number of panicle branches, 100-seed grain weight, and total grain weight. However, leaf width, leaf area, African rice gall midge (AFRGM) score, numbers of tillers and grain length were not significant. Factor analysis revealed that yield contribution factor was the most important characteristics, and its components include panicle length, number of spikelets, number of panicle branches, total grain weight, panicle weight, second iron toxicity score and panicle per meter square. The factor analysis characterized the traits into six factors, which accounted for 68 percent of the total variation. The cluster analysis classified the 25 upland rice cultivars into four groups based on characters measured. Cluster I consisted of 15 varieties and the traits whose average (within this class) were found to be higher than mean of all cultivars are considered superior to that overall cultivars for such character selection. These are leaf area, second iron toxicity score, African Rice Gall Midge (AFRGM) score and 100-seed grain weight. Therefore, this class could be used to grain quality selection. The second cluster consisted of 6 varieties, which had their averages higher than all cultivars for traits such as plant height, leaf length, and first iron toxicity score, panicle per square meter, number of tillers/stand, panicle length, number of spikelets, panicle weight, numbers of panicle branches, 100 seed grain weight and total grain weight. This class could be suitable for improving the yield components. Cluster III comprises of cluster II characters in addition to forage properties. Specific attribute of the each class could be used by breeders for rice improvement programs.