In the present study, a Box-Behnken experimental design was employed to statistically optimize the formulation parameters of chitosan phthalate and chitosan succinate microspheres preparation. These microspheres can be useful for oral insulin delivery system. The effects of three parameters namely polymer concentration, stirring speed and cross linking agent were studied. The fitted mathematical model allowed us to plot response surfaces curves and to determine optimal preparation conditions. Results clearly indicated that the crosslinking agent was the main factor influencing the insulin loading and releasing. The in vitro results indicated that chitosan succinate microspheres need high amount of crosslinking agent to control initial burst release compared to chitosan phthalate microspheres. The reason may be attributed that chitosan succinate is more hydrophilic than chitosan phthalate. The relative pharmacological efficacy for chitosan phthalate and chitosan succinate microspheres (18.66 +/- 3.84%, 16.24 +/- 4%) was almost three-fold higher than the efficacy of the oral insulin administration (4.68 +/- 1.52%). These findings suggest that these microspheres are promising carrier for oral insulin delivery system.