Present research deals with the comparison of neuro-fuzzy algorithms theoretical and application aspects. The compared algorithms are applied for a non-linear MISO system predictive model design. The modelled system is an unstudied continuous biotechnological process for protein biosynthesis. Simulative research is carried out in Fuzzy Toolbox of Matlab 5.2. Experimental approvement of the models is realized from experts in the Institute of Microbiology - Bulgarian Academy of Sciences.