The extraordinary conditions of a disaster, require the mobilisation of all available resources, inducing the rush of humanitarian partners into the affected area. This phenomenon called the proliferation of actors, causes serious problems during the disaster response phase including the oversupply, duplicated efforts, lack of planning. The aim of this research is to provide a solution to reduce the partner proliferation problem. To that end the main research question is put forward as “How to reduce the proliferation of partners in a disaster response”? Panel analysis of the historic record of 4,252 natural onset disasters between 1980 to 2013 via regression analysis, MA and AHP gives rise to the formation of a predictive decision-making platform called PREDIS. It is capable of predicting the human impact of the disaster (fatality, injured, homeless) of up to 3% of errors and enables the decision makers to estimate the required needs for each disaster and prioritises them based on the disaster type and socio-economics of the affected country. It further renders it possible to rank and optimise the desired partners based on the decision maker’s preferences. Verification of the PREDIS through a simulation game design using a sample group of decision makers, show that this technique enables the user to decide within one hour after the disaster strike using the widely available data at the time of the disaster. It also enables non-experts to decide almost identically to experts in terms of the similarity of the choices and the speed of the decision. The lack of an extensive database for the potential humanitarian partners from which to choose, is the limitation of this research in addition to the lack of standardised set of minimum requirements for the suitable partners. The model is also as strong as its data feed which is inconsistent in various humanitarian sources.