Protease is one of the most important industrial enzymes with a multitude of applications in both food and non-food sectors. Although most commercial proteases are microbial proteases, the potential of non-conventional protease sources, especially plants, should not be overlooked. In this study, horse mango (Mangifera foetida Lour) fruit, known to produce latex with a blistering effect upon contact with human skin, was chosen as a source of protease, and the effect of the extraction process on its protease activity evaluated. The crude enzyme was extracted from the kernels and extraction was optimized by a response surface methodology (RSM) using a central composite rotatable design (CCRD). The variables studied were pH (x(1)), CaCl(2) (x(2)), Triton X-100 (x(3)), and 1,4-dithryeitol (x(4)). The results obtained indicate that the quadratic model is significant for all the variables tested. Based on the RSM model generated, optimal extraction conditions were obtained at pH 6.0, 8.16 mM CaCl(2), 5.0% Triton X-100, and 10.0 mM DTT, and the estimated response was 95.5% (w/w). Verification test results showed that the difference between the calculated and the experimental protease activity value was only 2%. Based on the t-value, the effects of the variables arranged in ascending order of strength were CaCl(2) < pH < DTT < Triton X-100.