The queries on type-2 fuzzy relational databases may get non-reasonable answers, and as pointed out by Buckles, a tuple's membership value is not static but is a measure of the appropriateness of that tuple to a given query. Here, we use two supplementary attributes μ r and I e to model the concepts of fuzziness and uncertainty, respectively. We also use these two attribute values to compute the appropriateness of a tuple to a given query. In this paper, the proposed fuzzy relational model is defined by first-order logical. In our proposed model, the classical properties of the database, which are (1) database interpretation being redundant-free, being uniquely-determined and (2) the well-definedness of relational algebra, are preserved for the fuzzy relational algebra. Therefore, the proposed fuzzy relational model can capture more meaning than the conventional approaches and is more suitable to handle the real-life application.