Abstract:An ongoing study was directed to detect the best possible pretreatment process to improve the cellulose and reduce the total lignin yield in liquid ammonia (NH3)-pretreated bagasse. To examine the individual and cooperative effects of process variables: substrate concentration, NH3 loading, pretreatment temperature and time on the overall yield of cellulose, hemicellulose and lignin in resultant bagasse, statistical response surface methodology (RSM) design was applied. Tests were done by Box-Behnken Design (BBD) to set up the ideal pretreatment process conditions for bagasse. Results of the response surface design analysis indicate that ideal conditions to achieve the highest cellulose yield (44.6%) with decreased lignin yield (8.7%) at a substrate bagasse concentration is around 2%, NH3 loading is around 20%, treatment temperature is 120 ºC and retention time for 75 min. The good correlation noticed among the observed and predicted experimental results indicate the present model was useful to pretreat the lignocellulosic substrates. Model verification is specified, there is no disagreement among predicted and observed values. Higher reducing sugar (39.6 mg/g) yields were observed in high NH3 loaded treated bagasse than the low NH3 loaded pretreated bagasse (26.8 mg/g).