The paper is devoted to the development of railway infrastructure objects monitoring systems. The features of modern periodic and health monitoring sys- tems functioning are determined. It is proved that following the modern concept of threshold diagnostics causes low efficiency of monitoring systems functioning and the fact that the human factor becomes the determining influence on the results of diagnosis and prediction. The development of health monitoring should follow the path of intellectualization of data processing of diagnostic information and the in- tegration of health monitoring systems of infrastructure objects which are serviced by various linear departments of the railways. First of all, changes should happen at the stage of the conceptual notion about health monitoring systems, especially the requirements and measures for fixation of failure events and pre-failure conditions of diagnosed objects should be formulated on the basis of intelligent data mining. In the article basic definitions regulating the operation of monitoring systems are formulated, also the known definition of pre-failure condition is clarified. Moreo- ver, the existence of close frameworks of current instructions for maintenance and standards created on their basis for the development of monitoring systems using new principles of data processing is underlined. The consideration of monitoring systems peculiarities enables to increase its efficiency in the future in comparison with the implemented monitoring systems on the railways at the stage of design and construction of the monitoring system.