Thirty promising diverse genotypes of soybean were studied for genetic variability and genetic divergence during Kharif 2003. Genetic divergence using Mahalanobis D-2 statistic was studied in 30 genetically diverse genotypes of soybean for days to 50% flowering, days to maturity, plant height (cm), pod bearing length (cm), number of pod bearing nodes, number of seeds per plant, number of seeds per pod, hundred seed weight (g), number of seeds per plant, seed yield per plant (g), protein percentage (on moisture free basis) and oil content percent. Four characters like numb er of pods per plant, number of seeds per plant, pod bearing length and plant height were found as important sources of variation as observed from both cluster mean and variance analysis. The genotypes were grouped into eight different clusters. The cluster I was the biggest cluster consisted of eight genotypes followed by seven genotypes in cluster VIII, four in cluster III and three in cluster V with the minimum number of two genotypes in cluster II, IV, VI and VII respectively. The highest intra cluster distance was observed in cluster VI followed by cluster VIII, III and IV whereas cluster VII with minimum cluster distance. Thus, in the present study the better genotypes were identified in respect to most of characters on the basis of mean performance in the cluster, for oil percentage PK-416 and PS-10-24 were selected but for protein percentage the selected genotypes are PK-262 and SL-499. Genotypes viz. EC-34117, EC-398392, PS-10-24, DS-90-41, SL-517, PK-262, PS114 1, SL-499, PK-1162, PK-1283. PK-673. BS-9712, JS-81-13, MAUS-144, PLS-90, JS-82-180 and JS89-89 have been identified on the basis of high yielding ability.