The approach used to identify inbred lines that can produce superior hybrids is costly and time-consuming. It requires creation of all possible crosses and evaluation of the crosses to estimate combining abilities of parental lines for the desired traits. Predicting hybrid performance may help to reduce the number of crosses to be made and evaluated. The objective of this study was to determine whether traits measured on parental lines can be used to predict hybrid performance. Forty-three pollinator parents were crossed with three seed parents (AOK11, ATx3042 and ATx399) to develop 72 F1 hybrids. The hybrids and their parental lines were evaluated in two environments in a randomized complete block design with three replications. Data collected were days to 50 % flowering, plant height, thousand kernel weight and grain yield. Highly significant differences were observed for all traits in both parental lines and hybrids. The hybrids flowered earlier and were taller than their parental lines. Also, the hybrids had higher thousand kernel weight and grain yields than their parental lines.Significant correlations between the hybrid performance with the average performance of the parental lines were observed for all traits, ranging from r = 0.40 for grain yield to r = 0.65 for flowering time. Both mid-parent and better-parent heterosis were significantly correlated with the hybrid performance for all traits. Broad-sense heritability estimates for the traits ranged from (H2 = 0.19 for thousand kernel weight to 0.60 for flowering time) in the parental lines, and from 0.05 for thousand kernel weight to 0.57 for flowering time in the hybrids. However, the ability to predict hybrid performance using inbred line performance varied for different traits. These preliminary results show that studying parental inbred line performance could generate important information for predicting hybrid performance in sorghum.