The optimization of biochemical processes has drawn attention in recent years. The main objectives are production efficiency, cost reduction, improved product quality and time saving. This paper discusses an iterative closed-loop system identification algorithm to adjust a PID adaptive controller. The experiments on a fed-batch fermentation confirm that the model identification and the adaptive controller fit the biochemical reactors even under extremely variable conditions. On-line measurement of the relevant parameters, appropriate identification tools and intelligent adaptive controller appear to be efficient tools to control the non-linear dynamic bioreactors.