This article contributes a knowledge service model for knowledge-driven oil production decision. The core component of the model is the supporting environment construction for knowledge reasoning, which supports various assistant decision applications. The knowledge structure supporting relational data reasoning in the environment is proposed and it consists of three layers—meta-vocabulary layer, oil production fact layer, and decision strategy and design action layer. The formal definition and description of the knowledge structure is conducted. The evidence-production status reasoning mechanism is described. Following the mechanism, current oil production status is reasoned by analyzing the facts concealed in oil production information. The assistant decision system for oil-field dynamic development is developed by following the model.