This work deals with the optimization of the culture conditions of Bacillus invictae AH1 in order to increase the production level of the proteolytic activity. Response‐surface methodology (RSM) was applied for the most significant fermentation parameters (concentration of wheat bran and K2HPO4/KH2PO4) that were earlier identified by Plackett–Burman Design from seven possible factors. A central composite design was used and the quadratic regression model of producing active protease was built. A maximum protease activity was reached and validated experimentally, using a maximum wheat bran concentration (50 g/L) with increased K2HPO4/KH2PO4 concentration (2.275 g/L). Protease production obtained experimentally coincident with the predicted value and the model was proven to be adequate. Interestingly, the use of RSM increased the protease production by four times (7,000 U/mL) using a low‐cost substrate and a culture time of 40 hr, as compared to the standard culture conditions. In the second part of this study, a H2O2‐tolerant alkaline protease produced from B. invictae AH1 with a molecular mass of about 41 kDa, noted P3, was purified by successive steps of ultrafiltration, gel filtration and ion exchange chromatography. The K
m and Vmax values of the purified protease using casein, as substrate, were about 4 mg/mL and 27 μM/min, respectively. The highest enzyme activity was found at pH 9.0 and a temperature of 60°C. In addition, the enzyme showed a quasi‐total stability against H2O2 (5% for 1 hr) and against most of the tested solid and liquid detergents, suggesting its eventual use in bio‐detergent formulations.