This study set out to improve the physical and pharmaceutical characteristics of the present formulation using an appropriate experimental design. The work described here concerns the formulation of the dispersible tablet applying direct compression method containing roxithromycin in the form of coated granules. In this study $2^3$ factorial design was used as screening test model and Central Composite Design (CCC) associated with response surface methodology was used as optimization study model to develop and to optimize the proper formulation of roxithromycin dispersible tablet. The three independent variables investigated were functional excipients like binder (X1), disintegrant (X2) and lubricant (X3). The effects of these variables were investigated on the following responses: hardness (Y1), friability (Y2) and disintegration time (Y3) of tablet. Three replicates at the center levels of the each design were used to independently calculate the experimental error and to detect any curvature in the response surface. This enabled the best formulations to be selected objectively. The effect order of each term to all response variable was X3> X2> Xl> X1＊X2> X2＊X2> X2＊X3> X3＊X3> Xl＊X3> Xl＊Xl and model equations on each response variables were generated. Optimized compositions of formula were accordingly computed using those model equations and confirmed by following demonstration study. As a result, this study has demonstrated the efficiency and effectiveness of using a systematic formulation optimization process to develop the tablet formulation of roxithromycin dispersible tablet with limited experiment.