In this study we examined the differential contribution of superior parietal cortex (SPC) and caudal dorsal–lateral prefrontal cortex (dlPFC) to drawing geometrical shapes. Monkeys were trained to draw triangles, squares, trapezoids and inverted triangles while we recorded the activity of small ensembles of neurons in caudal area 46 and areas 5 and 2 of parietal cortex. We analyzed the drawing factors encoded by individual neurons by fitting a step-wise general-linear model using as our dependent variable the firing rate averaged over segments of the produced trajectories. This analysis demonstrated that both cognitive (shape and segment serial position) and motor (maximum speed, position and direction of segment) factors modulated the activity of individual neurons. Furthermore, SPC had an enriched representation of both shape and motor factors, with the motor enrichment being stronger than the shape enrichment.Following this we used the activity in the simultaneously recorded neural ensembles to predict the hand velocity. In these analyses we found that the prediction of the hand velocity was better when we estimated different linear decoding functions for each shape than when we estimated a single function across shapes, although it was a subtle effect. Furthermore, we also found that ensembles of caudal dlPFC neurons carried considerable information about hand velocity, a purely motor factor. However, the SPC ensembles carried more information at the ensemble level as a function of the ensemble size than the caudal dlPFC ensembles, although the differences were not dramatic. Finally, an analysis of the response latencies of individual neurons showed that the caudal dlPFC representation was more sensory than the SPC representation, which was equally sensory and motor. Thus, this neurophysiological evidence suggests that both SPC and caudal dlPFC have a role in drawing, but that SPC plays a larger role in both the cognitive and the motor components.