A major difficulty in the monitoring and control of bioprocesses is the lack of reliable and simple sensors for following the evolution of the main state variables and parameters such as biomass, substrate, product and growth rate. An adaptive estimation algorithm is proposed to recover the state and parameters in bioprocesses. This estimator utilises the physical process model and the reference model approach. Experiments concerning estimation of biomass and product concentrations and specific growth rate, during batch, fed-batch and continuous fermentation processes are presented. The results show the performance of this adaptive estimation approach.