In recent years a remarkable effort has been made in the development of new sensors and process analytical technology. However, it is still difficult to find reliable and low cost commercial sensors for on-line measurements of important variables. Therefore, considerable attention has been focused on the development of on-line software sensors. Nevertheless, the application of those algorithms to complex biological processes is still very incipient. In this work two different state estimators have been studied regarding their applicability to the recombinant Escherichia coli fed-batch fermentation. Both algorithms showed the ability to estimate on-line biomass and acetate concentrations. However, the extended Kalman observer exhibited a better convergence in spite of being less flexible regarding the combination of the measured and estimated variables. Copyright © 2005 IFAC