Concept lattice is an important method and technology for data analysis and knowledge discovery, however, in this research field, many researchers focus on two-valued formal context. In this paper, we transform continuous-valued formal context into many-valued formal context first, then on the basis of many-valued formal context, the definition of the equivalence class of single attribute and formal pairs of single attribute are given. We discuss constructing formal concept lattice based on single attribute concepts, uniting concepts, updating concept and adding concepts, all these concepts are generated by union, intersection and equality of objects of formal pairs. At last, we present a fast algorithm for constructing concept lattice and analyze its complexity. It has demonstrated that the algorithm is effective by our experiments.