Osmotic stress (OS) techniques have found wide application in the crystallization of proteins, the study of intermolecular forces between biocolloids, and in studies of the role of solvent in molecular recognition events. In these cases, the role of the solvent is not well understood. We present here a combined Monte Carlo simulation and smoothed density functional theory approach with which model colloid-solvent mixtures may be studied in a potential of mean force approximation. In particular, our simulations focus on parallel cube colloids in Lennard-Jones solvents. We have applied this approach to simulate some OS experiments in which the dependence of the intermolecular separation (h) of DNA strands in aqueous solutions on an applied osmotic pressure (II) have been measured. Our simulations demonstrate that an underlying oscillatory solvation potential qualitatively produces the features of II-h relationships observed in the experiments. In particular, increasing II produces a transition to a more dense colloidal crystal. More importantly, the transition II decreases with increasing temperature as observed in experiments. In addition, hydrostatic pressure p(b) and II are shown to operate conversely in isothermal solvent controlled OS experiments. In addition to characterizing some key features of solvent controlled processes, these simulations reconcile OS experiments with surface forces apparatus (SFA) experiments, and suggest new experimental routes to studying intermolecular forces.