Peer-to-Peer eCommerce communities are commonly perceived as an environment offering both opportunities and threats. One way to minimize threats in such an open community is to use reputations to help evaluating the trustworthiness and predicting the future behaviour of peers. The subjective nature of trust results in its uncertainty and fuzziness characters. The present paper introduces a formalism model of subjective trust by which we can transform between qualitative reputation and quantitative voting data. We also bring forward algorithms to compute direct trust and recommender trust and propose a similarity measuring method which can be used to compare the similarity of two users? reputation on knowledge level. The present model properly settles the uncertainty and fuzziness properties of subjective trust which is always the weakness of traditional subjective trust model, and provides a step in the direction of proper understanding and definition of human trust.