Taking some steam flooding experimental area in the western depression of the Liaohe Basin in China as an example, it quantitatively evaluates the flow units of the sedimentary facies in fan delta front and analyzes the influences of the flow units on the development of the oilfields. Comprehensive analysis of multifarious data such as geology, core analysis and well logs indicates that porosity, permeability and FZI are key parameters for classification and evaluation of the reservoir flow units. Then using SPSS cluster analysis software platform, cluster analysis are conducted on reservoir physical properties of cored wells, and based on which discriminant analysis are performed to obtain the discriminant formula. The discriminant analysis of reservoir physical properties of the wells without core data is based on fine interpretation of well logging data. Combined with the results of cored wells, the classification and evaluation of the flow units has been completed in the target layer of the whole study area. Finally the flow units are divided into four classes, namely, Class A, Class B, Class C and Class D. Classes A and B flow units represent high-quality reservoirs. The flow units are mainly controlled by sedimentation. Classes A and B are mainly sedimentary microfacies of underwater distributary channel and distributary mouth sand bar. The paper proposes five principles of judging the rationality of flow units, describes the development characteristics of different classes of flow units on section and plane, analyzes the influence of flow units on oilfield development and suggests countermeasures. Specially, in the process of steam flooding, target injection layers should be optimized and steam injection pressure should be effectively controlled to prevent steam channeling; when planning infill wells, injection and production wells should be deployed in the same class of flow unit to ensure good connectivity; the development of remaining oil should focus on Classes A and B flow units, and meanwhile pay more attention to the fine interpretation results of well logging data.