In this study, a central composite rotatable design based on response surface methodology (RSM) was employed to design and formulate an appropriate paeonol microparticle formulation. Five levels of a three-factor, rotatable, central composite design were used to evaluate the critical formulation variables. The optimum conditions for preparing paeonolloaded microparticles were predicted to be: polyvinyl alcohol (PVA) content (2.84%), the ratio of drug to polymer (6.88) and the stirring rate (1007.59 rpm). The optimized responses for production yield and loading efficiency were found to be 68.86% and 55.90%, respectively, and the particle size were 23.27 +/- 0.76 mu m and the sorting coefficient (sigma) was 0.732. Furthermore, in vitro release study suggested that microparticle could be a suitable delivery system in treating skin disease for its sustained release of drug. In conclusion, RSM can be successfully used to optimize the effect of formulation variables.