Abstract: Predictive Results of oil production by polymer flooding process are directly related to annual planning and mid-long term planning deployment in the process of oil field development, so the selection of forecast model is particularly important. This paper briefly introduces the combination forecast model and develops an optimal combination model for oil production forecast based on Weibull, HCZ and generalized Weng's prediction models, in which the weighting coefficient is determined by the criterion of minimum square sum of predictive error. The precision of such a model is compared with those given by the three models by means of real production data, proving this model is effective and practicable.

Key words: Karamay oil field, combination, polymer flooding, oil production prediction, weighting coefficient