Abstract: In order to accurately predict the energy consumption in the oil and gas production process， the primary energy consumption indicator was determined through the analysis of the oil and gas production process. A hybrid modeling method was proposed， in which the mechanism model was combined with LSSVM to predict the comprehensive energy consumption of oil and gas production process. Based on the analysis of the whole production process and the relationship between subprocesses， the mechanistic model of the energy consumption was established， and then LSSVM was applied to compensate the error which could not be described using the mechanism model. The simulation results showed that the energy consumption of oil and gas production process could be accurately predicted with the proposed method， and the performance was better than those of mechanistic model and the model established by LSSVM. It could be also concluded that the hybrid model has better stability and reliability. When the model was applied to the actual production process in some oil recovery operation area， satisfactory results could be received.

Key words: oil and gas production, energy consumption indicator, mechanism model, LSSVM, hybrid model