The way of using optimal control theory to implement the optimal fuel control strategy of hybrid excavator is limited, because its computational complexity is large and the working conditions must be known in advance for global optimization. A real-time optimal fuel control strategy was proposed in this paper to solve the problem. An engine model of "speed-power-fuel consumption rate" was established. Under the constraint of DC bus voltage stability, the power compensation of the energy storage system was calculated as a control instruction, which can make the engine work efficiently. Finite control set model predictive control algorithm was proposed to follow the instruction speedily and flexibly. Through the simulation, the effectiveness of the proposed approach was demonstrated. Engineering practice results indicate that the fuel consumption is 82.2% and 77.6% of the prediction of the traditional model with flat light load and heavy load, respectively.