Consumption of pulses is associated with many health benefits by mechanisms that are not fully understood. This study sought to identify the starch component(s) in cooked lentils responsible for lowering postprandial glycemic response (PPGR). Rapidly digestible (RDS), slowly digestible (SDS) and resistant starch (RS) content of 20 varieties of cooked lentil were determined by in vitro methods and 8 varieties, representing a linear range of SDS, were chosen for a human trial with 10 participants to determine PPGR and glycemic index (GI). Among the 20 lentil varieties, RS accounted for 35% of the variation of in vitro area under the starch hydrolysis curve (SHAUC) (r = -0.587; p < 0.01), but RDS (r = 0.401; p = 0.080) and SDS (r = -0.022; p = 0.927) were not significantly related to SHAUC. Multiple linear regression of in vitro data resulted in an equation [SHAUCest = 30.9RDS - 63.6RS + 9680] that accounted for 70% of the variance in SHAUC, with SDS excluded due to collinearity. In the human trial all 8 lentils had low GI values (10 to 23). Neither GI nor area under the glucose response curve (AUC) was significantly related to RDS, SDS or RS (minimum p = 0.24). However, SHAUC and SHAUCest, respectively, were related to both GI (r = 0.704, p = 0.051; r = 0.773, p = 0.024) and AUC (r = 0.765, p = 0.027; r = 0.822, p = 0.012). These results confirm that lentils have low GI values, which is not reliably predicted by their RDS, SDS and RS contents when considered individually. However, in vitro SHAUC and a combination of RDS and RS may be predictive of the PPGR of lentils.