This work aims to present a methodology based on a proactive approach for the early detection of possible liquids carryover towards compressors installed in upstream process plants where the presence of heavy hydrocarbon inside gas stream is a recursive problem. Analyzing trough remote monitoring some parameters of the facilities that are upstream to the compressor and, simultaneously, some machine critical instrumentation that was selected through early detection algorithms, it is possible to develop real time synthetic indicators that warn the possibility of liquid ingestion. An in-depth analysis of the plant and of the process lines and facilities that are upstream to the compressors, permits to identify the critical parameters to monitor, which unequivocally indicate a change in both physical conditions and gas composition. This methodology has been applied to create tailor made synthetic indicators that have been implemented in operating sites, to highlight the upset in the process parameters that could have led to liquid carry over with consequent possible mechanical breakdowns. These indicators enhance the deep analysis of these phenomena, resulting in the development of tailored operative adjustment to be applied at the early stage of the liquid carryover occurrence. The combined monitoring both on the upstream facilities and on the on-board instrumentation of compressors allows to have an early detection of the liquid carryover events, giving time to the operators to understand the situation and act accordingly to avoid hazardous conditions. The methodology innovative feature is the ability to monitor, by remote, a combination of critical parameters at "plant level", together with cross-check trough machine-learning algorithms of the relevant on-board instrumentation at "compressor level" to have a complete picture of the phenomena and prevent the compressors from unplanned damages due to liquid carryover.