Central composite design of response surface methodology (RSM) was employed to optimize Lactobacillus content (X₁: 4.5 × 10⁶ to 5.5 × 10⁶ cfu/g), Bacillus subtilis content (X₂: 9.0 × 10⁶ to 1.1 × 10⁷ cfu/g) and yeast content (X₃: 5.50 × 10⁶ to 7.0 × 10⁶ cfu/g) of solid-state fermentation (SSF) cottonseed meal with low pH value, high dry matter recovery (DMR) and high the reducing-sugar content. Results indicated that the data were adequately fitted into 3 s-order polynomial models. The Lactobacillus content, Bacillus subtilis content, and yeast content were found to have significant linear, quadratic and interaction effects on pH value, the DMR and the reducing-sugar. The optimal extraction conditions were predicted to be lactobacillus content of 5.50 × 10⁶ cfu/g, Bacillus subtilis content of 1.08 × 10⁷ cfu/g and yeast content of 6.08 × 10⁵ cfu/g. The pH value, DMR and the reducing sugar predicted by RSA were 5.01, 91.8% and 1.69%, respectively. The detection index obtained experimentally was close to its predicted values. The establishment of such model provides a good experimental basis employing RSM for optimizing the quantity of inoculation of Lactobacillus, Bacillus subtilis, and yeast on fermentation.