This paper has developed a recursive least squares scheme for operating a class of continuous fermentation processes at the optimal steady state productivity. More precisely, the class of continuous fermentation processes under investigation is described by a widely accepted non-segregated fermentation process model. Based on a recursive least squares algorithm, an on-line estimation scheme for estimating the optimal set points for feed substrate concentration and dilution rate is developed. Finally, via simulation, it is shown that this on-line estimation scheme is robust against measurement time (sampling time).