Computational methods that predict tissue assembly aid in the production of biological substitutes that mimic native tissue. In particular prostate cancer cells self-assemble on an attachment-limiting substrate into spheroids that resemble micrometastases and have application to in vitro drug testing. Our research group has developed mathematical models that predict cell aggregation kinetics in this culture system. Through an integrative experimental and computational approach, we are refining our models to account for the restructuring dynamics by which 2D aggregates form 3D multicellular spheroids. Our study investigates the mechanism of restructuring and the roles of cell adhesion and motility in this process for DU 145 and LNCaP prostate cancer cells. Irregularly shaped, 2D aggregates restructure through incremental cell movements into 3D spheroids that we have simulated with Monte Carlo techniques. Of the two cultures examined, restructuring is more pronounced for DU 145 aggregates. Motile DU 145 cells form spheroids with a minimum cell overlay of 30% for 25-mers as estimated by simulation vs. 5% for adhesive LNCaP cells in aggregates of the same size. Over 72 h, the texture ratio increases from 0.55 ± 0.05 for DU 145 aggregates with projected areas exceeding 2000 sq. microns to a value approaching 0.75 ± 0.02 (p < 0.05). For LNCaP aggregates of comparable size, the increase in texture ratio is more modest, less than 15% during the same time period (p < 0.05). Combined these data suggest that motility events, rather than adhesion, govern the overall rate of spheroid restructuring. Since spheroid composition is size-dependent, our models may be able to predict both spheroid size and composition from properties of the inoculum.