In this paper, we describe an application of Bayesian networks in the prognoses of the properly utilization of thermostatic valves, by using the SamIam program. The application was validated based on the data obtained by measures in the District Heating System of Bucharest. Thermostatic valves transformed heating systems into efficient technical facilities, able to properly react to any internal and external disturbing factors. The diagnoses concerning the properly use of thermostatic valves represent the original contribution of the research team. At the end of the paper it will be illustrated practical applications of the theoretical knowledge discussed throughout the article in a number of case studies which present several problems related to the thermostatic valves.