The experiment was conducted during kharif 2010-2011 with 41 genotypes of soybean to study genetic variability, correlation and path coefficient analysis of yield and its component characters. In general, PCV were higher than the GCV values for all the characters. The broad sense heritability estimates were high for all the 9 characters. Correlation and path coefficient analysis for seed yield/plant and its components indicated that out of 9 characters, seed yield/plant showed positive and significant association with biological yield, pods per plant and 100-seed weight indicating that an intense selection for these characters will improve seed yield in soybean. Path coefficient analysis revealed that biological yield, number of pods per plant and 100 seed weight were major characters influencing seed yield directly and indirectly. The results indicated that biological yield is responsible for manipulation of seed yield in soybean.