In glutamic acid fermentation, the molasses feeding policy and time of penicillin addition significantly affected glutamic acid production, and a fuzzy supervisory control system was developed for their quasi-optimal regulation. From the trend of the experimental data, production rules and membership functions of fuzzy inference were devised to determine the quasi-optimum molasses feeding policy and penicillin addition time. A computer with a multitasking operating system was used for the construction of the control system with fuzzy inferencing, which decided the control policy every minute, and the feed rate was controlled automatically. The pattern of residual sugar concentration was almost the same as that of maximum glutamic acid production under manual operation. Using the computer control system, stable production was maintained at the highest level of 71 to 75 g/L.