An optimization strategy, based on statistical experimental design, is employed to enhance the production of thermostable α-amylase by a thermotolerant B. licheniformis AI20 isolate. Using one variant at time (OVAT) method, starch, yeast extract, and CaCl2 were observed to influence the enzyme production significantly. Thereafter, the response surface methodology (RSM) was adopted to acquire the best process conditions among the selected variables, where a three-level Box-Behnken design was employed to create a polynomial quadratic model correlating the relationship between the three variables and α-amylase activity. The optimal combination of the major constituents of media for α-amylase production was 1.0% starch, 0.75% yeast extract, and 0.02% CaCl2. The predicted optimum α-amylase activity was 384 U/mL/min, which is two folds more than the basal medium conditions. The produced α-amylase was purified through various chromatographic techniques. The estimated enzyme molecular mass was 55 kDa and the α-amylase had an optimal temperature and pH of 60–80°C and 6–7.5, respectively. Values of and for the purified enzyme were 454 mU/mg and 0.709 mg/mL. The α-amylase enzyme showed great stability against different solvents. Additionally, the enzyme activity was slightly inhibited by detergents, sodium dodecyl sulphate (SDS), or chelating agents such as EDTA and EGTA. On the other hand, great enzyme stability against different divalent metal ions was observed at 0.1 mM concentration, but 10 mM of Cu2+ or Zn2+ reduced the enzyme activity by 25 and 55%, respectively.