With the services popularity increase and providers number expansion, several web services fulfilling the desired task are available and functional characteristics prove their limits as a selection criteria. To cope with this, many researchers propose architectures and approaches to integrate non functional features into web services selection process and some of them adopt user preferences centric strategies. We present in this paper a new system for scoring, ranking and classifying web services by their relevancy to well-expressed non functional requirements (NFRs). The ultimate goal is to assist the user specifying his NFRs, compute scores of candidate web services and provide the rank and class of each to rapidly and optimally recommend the best web service. Experimental results confirm the efficiency of the proposed algorithms and the accuracy of obtained results.