This study enhanced the production efficiency of Lacticaseibacillus casei TCS by optimizing medium composition and fermentation conditions for high-density culture. Initially, single-factor and orthogonal experimental designs and the response surface methodology were used to determine the optimal concentrations of medium. Subsequently, we applied the artificial neural network-genetic algorithm optimization method, which significantly increased the viable bacterial count. Fermentation kinetics were modeled using logistic growth and Luedeking-Piret models, which accurately predicted cell growth. Amberlite IRA 67, an anion exchange resin, effectively adsorbed lactic acid and maintained pH levels. Furthermore, the combined use of fed-batch fermentation and ion exchange alleviated the effects of acid inhibition, salt stress, and substrate limitation, resulting in a maximum cell density of 10.01 lg CFU/mL, a 9.3-fold increase over the basal medium. This study develops a robust and cost-effective strategy for the industrial production of