In the present study, compactin production by Penicillium brevicompactum WA 2315 was optimized using solid-state fermentation. The initial one factor at a time approach resulted in improved compactin production of 905 microg gds(-1) compared to initial 450 microg gds(-1). Subsequently, nutritional, physiological, and biological parameters were screened using fractional factorial and Box-Behnken design. The fractional factorial design studied inoculum age, inoculum volume, pH, NaCl, NH(4)NO(3), MgSO(4), and KH(2)PO(4). All parameters were found to be significant except pH and KH(2)PO(4). The Box-Behnken design studied inoculum volume, inoculum age, glycerol, and NH(4)NO(3) at three different levels. Inoculum volume (p = 0.0013) and glycerol (p = 0.0001) were significant factors with greater effect on response. The interaction effects were not significant. The validation study using model-defined conditions resulted in an improved yield of 1,250 microg gds(-1) compactin. Further improvement in yield was obtained using fed batch mode of carbon supplementation. The feeding of glycerol (20% v/v) on day 3 resulted in further improved compactin yield of 1,406 microg gds(-1). The present study demonstrates that agro-industrial residues can be successfully used for compactin production, and statistical experiment designs provide an easy tool to improve the process conditions for secondary metabolite production.