Abstract: To address the issue of high energy consumption in drilling operations of electric drill jumbos, this paper proposes a hybrid energy consumption prediction model that integrates physical mechanisms with data-driven methods. The energy transfer equation and an Amesim simulation model of the hydraulic rock drill are established, and the effects of pressure parameters and rock properties on energy consumption are analyzed through rock drilling experiments. The results show that increasing the impact pressure can accelerate the penetration rate but simultaneously leads to higher energy consumption. Insufficient feed pressure causes idle striking, while excessively high damping pressure alters the force distribution on the drill tool. The response pattern of energy consumption is governed by rock strength and wave impedance characteristics, with granite exhibiting significantly higher energy consumption than sandstone and limestone. Based on the simulation model, a least squares support vector machine (LSSVM) is introduced to construct a data compensation module. To enhance the overall model performance, a sine cosine algorithm (SCA) is employed to automatically optimize the hyperparameters of LSSVM, ultimately forming a hybrid-driven prediction model. Experimental results demonstrate that the proposed hybrid model achieves a mean absolute percentage error (MAPE) of 1.61%, which is 2.69% and 5.58% lower than that of the standalone LSSVM model and the pure physical model, respectively, substantially improving the accuracy and reliability of energy consumption prediction.