Experimental design and surface response methodology were used to study the production of radish oil ethyl ester, using Burkholderia cepacia lipase. The estimative of the standard effects were determined as a function of time to analyze the importance of some variables. The results showed the importance of variables, as well the extents of their effect, vary throughout the reaction. Then was possible to analyze the behavior of the reaction medium and to propose situations where higher yields could be obtained. In fact, when the reactor was operated in this way, an 80,1% yield was achieved, much higher than common operation (39.9%). Also, enzymatic activity of ester production was tenfold.