Abstract: Traditionally the screening of viscous crude oil production schemes are conducted manually, in which every experts may arrive at different results and the superiority of these are hard to judge. The mixed intelligence system which is more reliable is put forward. In the system, the symbolic expert system (SBWS) and the artificial neural network are integrated, the former being aimed at optimizing the structure of the latter while the latter being used to acquire the contribution degrees of screening of viscous crude oil production schemes’ character parameters. With the system, fuzzy expression of expert knowledge, multi-level organization of the knowledge base, fuzzy inference are realized. A satisfactory result was yielded in the field testing. The development of the system also opens a new perspective toward a more extensive utilization of the mixed intelligence system in other areas of oil exploration and development.