Different from explicit customer needs that can be identified directly by analyzing raw data from the customers, latent customer needs are often implied in the semantics of use cases underlying customer needs information. Due to difficulties in understanding semantic implications associated with use cases, typical text mining-based methods can hardly identify latent customer needs, as opposite to keywords mining for explicit customer needs. This paper proposes a two-layer model for latent customer needs elicitation through use case reasoning. The first layer emphasizes sentiment analysis, aiming to identify explicit customer needs based on the product attributes and ordinary use cases extracted from online product reviews. Fuzzy support vector machines (SVMs) are developed to build sentiment prediction models based on a list of affective lexicons. The second layer is geared toward use case analogical reasoning, to identify implicit characteristics of latent customer needs by reasoning the semantic similarities and differences analogically between the ordinary and extraordinary use cases. Case-based reasoning (CBR) is utilized to perform case retrieval and case adaptation. A case study of Kindle Fire HD 7 in. tablet is developed to illustrate the potential and feasibility of the proposed method.