In the modern complex industrial process, the control system generally has characterstics of large intertila, nonlinearity and time-varyting, and isi\\ts control requirements are diverse and uncertain, so it is diffcult to smoothly trun the control parameters. To solve the problem, fuzzy evaluation approach is used to improve the SGA (simple genetic algorithms), and a fuzzy finess function is designed to divide those control requirements into many evaluation factors with different weights. The individual in GA (gentice algorithms) is control parmeters. The fintness of the individual reflects the fuzzy evaluation degree of control result, and shows the approximate degree of control result and ideal situation. In thd paper, we use the fuzzy-GA to optimixe the control parmeters of temperature controller in tower type fermenter, Experiments and simulations show that control indexes have been improved and this approach can saccessfally solve parmeter optimization problem in complex industrial process.