Introduction: In the recent years, application of microorganisms as green biocatalysts for removing caffeine pollution from industrial wastes and food caffeinated have been extensively considered. This investigation reports on optimization of bio-decaffeination process under growing cells of Saccharomyces cerevisiae by using the Taguchi statistical approach.  Materials and methods: Five variables, i.e. caffeine, Zn+2, glucose, peptone concentrations and time incubation, which have significant effects on bio-decaffeination process, were selected and L16 (44× 13) orthogonal array was determined for experimental trials. Caffeine degradation was estimated by HPLC (High Performance Liquid Chromatography) analysis. Results: Use of Taguchi approach for optimization of design parameters resulted in about 82.8 % reduction of caffeine in 48 h incubation when 3g/l peptone, 5mM Zn+2 ion and 5 g/l of caffeine are present in the designed media. Under the optimized conditions, the yield of degradation of caffeine (5 g/l) by the growing cells of yeast strain TFS9 has been increased from 25.5 to 82.8 % which is 3.2 fold higher than the normal yield. The improvement of caffeine removal after best conditions were made shows the efficiency of Taguchi experimental design in such studies. Discussion and conclusion: The current investigation is the first report for successful application of the Taguchi experimental approach to the bio-decaffeination process. According to the analysis of experimental results, the present study proposes the potentiality of the Taguchi approach to enhance the bio-decaffeination performance with the native strain of Saccharomyces cerevisiae.