The aim of the present work was to optimize a tablet formulation containing a physical mixture of a practically insoluble drug (prednisone) with a superdisintegrant (croscarmellose sodium) and two filler-binders characterized by differing water solubility (dicalcium phosphate dihydrate and anhydrous beta-lactose). Crushing strength, disintegration, and dissolution were measured for 10 formulations distributed over a factor space according to a simplex lattice design for a special cubic model. Multiple linear regression analysis was used to assess the best fit for each variable. The model predicted that increasing the amount of disintegrant to a critical amount (50%) would result in reduced disintegration time for dicalcium phosphate/beta-lactose ratios > 0.3, no changes in disintegration time for ratios < 0.3, and for all ratios an improvement in dissolution at 10 min. Crushing strength values of dicalcium phosphate increased with increasing disintegration concentration but not for beta-lactose tablets. The physical mixture of a practically insoluble drug with a superdisintegrant was confirmed as a valid approach to the improvement of dissolution, even in presence of other components. The solubility of the filler-binders influenced the minimum amount of disintegrant needed; when a soluble diluent was used, the amount of disintegrant required was reduced.