The purpose of this work is to obtain alkaline protease overproducing strain Bacillus subtilis TISTR 25 by using UV light and NTG as mutagens. Potency Index was used for primary screening. The rapid primary screening for alkaline protease overproducing mutants from the parental strain on skim milk plate are typical for identification of each mutant by halo formation. After that secondary screening was done by determination of alkaline protease activity in fermentation broth by casinolysis method. This experiment showed that the correlation between potency Index and alkaline protease activity was linear. Mutagenesis was performed using the two consecutive 20 seconds irradiation time at wavelength 254 nm and a distance 20 cm from the UV lamp. The survival rate was found to be 0.1%. The two consecutive NTG concentration for the induction was 25 ug/ml. A mutant named “UUNN-l” was selected producing alkaline protease at 81.80% higher than that produced by Bacillus subtilis TISTR 25. Cultivation of UUNN-l in shaking flask, the alkaline protease activity was at pH at 9.5 and showed optimal temperature of 45 ℃ and optimal pH of 9.5 the latter differed from the enzyme produced by Bacillus subtilis TISTR 25. Cultivation in 5L fermenter containing the mixture of soy bean meal and sunflower meal and 0.3 g% as nitrogen source and 0.5% cassava starch as carbon source gave maximum alkaline protease activity of 640.15 units/ml after 60 hours. The UUNN-1 broth was precipitated with 70% ammonium sulfate and over dried at 45 ℃ . The alkaline protease powder had the specific activity of 709.91 units/mg protein, 85.25% higher than that produced by Bacillus subtilis TISTR 25. The alkaline protease powder extracted from UUNN-1 broth was stable below 37 ℃ within 60 days.