<p>Today in the context of Cloud Computing, given the variation in consumer needs and the large number of suppliers, it is necessary to encourage these suppliers to collect and seriously study customer needs, in order to constantly improve their services and therefore attract more consumers. This necessarily calls for the use of the requirements engineering process and particularly an appropriate elicitation approach. To reply to this problem, we propose an elicitation approach based on the similarity of appearance between needs in the Cloud context, the distance, and the state of the consumer need. This approach makes it possible to know the existing needs and services, to classify these needs, and to elucidate exactly the most common and similar ones. The expected result of this approach is to have groups of needs classified according to the state of the need suggested, used, or requested according to the consumer requirement on this need, and if it is priority or preferred. Finally, we evaluate the performance of the proposed approach by applying accuracy and precision metrics, which yield promising results.</p>