由于受沉积环境的影响，低渗透油气田储层砂体连通性差，渗透率低，油水流动阻力大，生产规律完全不同于常规油气藏，生产中油气产量变化极其复杂，预测难度大。介绍了双对数模型、威布尔模型、俞启泰模型、灰色系统模型和微分模拟模型等5种目前进行油气产量预测的主要模型，以渤海湾盆地高5块低渗透油气田为研究对象，利用生产动态数据，运用5种模型分别对研究区产量进行了预测和分析，对比了各种方法在油气田不同开发阶段的预测精度，指出灰色系统模型对预测产量的单调上升或单调下降较为准确，而其他模型则可用于对油田开发指标的全程预测。但对渤海湾盆地高5块而言，5种常用模型的预测都存大较大的误差，为此重新拟合出了一种精度较高的线型预测模型。

The low permeability petroleum reservoir, which is controlled by depositional environment, is characterized as low permeability and bad connectivity between sand bodies, associated with big resistance of fluid.  The petroleum product pattern for the low permeability reservoir is different with normal permeability reservoir, causing the difficult prediction of petroleum production. In this paper, we introduced 5 predicted production patterns (i.e. double log predict, Weibull, Yuqitai, gray system and fluxionary simulation) to predict the petroleum production in the Gao 5 block. The comparative results suggest that the grey system would have an ability to predict the simple change of petroleum product whereas others provide the complex change. But, there is big error of petroleum product production by using of these five patterns. So we developed a linear pattern of petroleum production prediction that can provide the better prediction of petroleum production.