Abstract: he operation cost holds a great proportion in oil and gas production，directly affecting economical benefit of oil and gas exploitation.Accurate prediction makes effective cost control possible and hence increase corporate benefits.This paper applies the combination forecasting method in the operation cost prediction，combining the nonlinear regression model and the grey model with proper weight.A comparison of mean absolute error (MAE)，mean average absolute percentage error (MAPE)，mean square error (MSE)，and mean square percentage error (MSPE) proves that combination forecasting model works better than single models，and is able to improve precision of oil and gas operation cost prediction.

Key words: operation cost, combination forecasting model, regression model, grey model