In the first part of this paper two solutions for the temperature automatic control of the fermentation process are proposed. The first adopted solution control used a PI controller, for which the parameters were established in order to avoid the actuator saturation that adjusts the feeder flow with cooling agent. The second solution was consisted in using the logic' fuzzy and had as aim to get some good properties of ruggedness, general recognise, of the fuzzy command, at the dynamic properties change and in the conditions of some non-linearity such is the control valve saturation. Was emphasised that the two different methods lead to the same results. The second part of the article presents a comparative analysis of the possibilities offered by two approaches regarding the indirect way of collecting the information about the process by neural modelling.