Fractured biological limestone reservoirs are characterized by structural complexity, strong inhomogeneity, and dual pore networks,resulting in a lack of technical methods for fracture identification for biological limestone, so it is a big problem to recognize and evaluate the fracture development of this type of reservoir. In this paper, taking W oilfield as an example, based on various data such as core data, CT scanning, injecting thin sections conventional logging curves and FMI imaging, fracture reservoir prediction models were constructed: through the comparison of the core photo, thin section and CT, the "micro" fracture was identified, then the response characteristics of dual laterolog for different production fractures were summarized and microfractures could be found in the biological limestone phase of the research section according to the large difference in the amplitude of the electrical resistivity; by constructing a BP artificial neural network, the single-well fractured reservoirs were delineated; by the comprehensive probabilistic index approach to realize the prediction of the distribution of cracked reservoirs in the whole region. The characteristics of oil and gas production at the fracture development and non-development places were compared and summarized, which provides theoretical basis for the research on the relationship between fracture development and oil and gas production.