Cooked whole pulses are widely accepted as a low glycemic food with potential health benefits of reducing the risk of obesity and type 2 diabetes. However, the structural basis for the variation in digestion rate and extent of different pulses is still unclear. In this study, intact cotyledon cells of four commercial pulses (i.e., pinto bean, garbanzo bean, green-split pea, and black-eyed pea) were isolated under controlled cooking and isolation conditions, as a model for whole foods. We investigated the macrostructure of intact cells and structural features of entrapped starch granules, such as crystallinity and thermal parameters, and diffusion of an amylase-sized FITC dextran probe into intact cells. In vitro digestion kinetic profiles of pulse cells were monitored by a reducing sugar assay with a fixed α-amylase activity, and were fitted into a first order model to obtain the apparent rate coefficient and digestion extent at 180 min. Pearson's correlation analysis suggested that there is no significant correlation between the kinetic parameters and structural features of entrapped starch granules nor the size of the cell particles. It is concluded that the starch digestibility of isolated pulses is controlled by the accessibility of digestive enzymes, which is limited by the rigid cell wall and/or pulse protein matrix, rather than starch structure.