Abstract To solve the problem of poor fracture identifying effect on electrical logging in oil-based mud, the application of acoustic logging to the quantitative characterisation of fractures is expanded from three aspects, namely, Stoneley waves, longitudinal and transverse waves and cross dipole acoustic waves, and a fracture logging evaluation model closely related to production capacity is established considering the radial extension characteristics of fractures. The Stoneley reflection coefficient is used to determine fractures locations to help detect fractures during normal micro-resistivity imaging logging. Based on the experiment on the relationship between fracture width and acoustic attenuation coefficient, empirical formulae for calculating fracture width have been established by primary wave and shear wave energy information considering the effect of porosity. The new parameters, including spectrum correlation coefficient and energy difference from cross dipole array acoustic logging data, can be used for fractures evaluation. The more developed the fractures are, the greater the energy difference becomes, and the smaller the spectrum correlation coefficient is, the higher the production is. The fracture effective evaluation parameters can be separated into two components, specified as the degree of fracture vertical opening and radial extension. Combining the conventional logging and array acoustic logging (including cross dipole array acoustic logging), a fracture radial extension evaluation model is presented closely related to productivity.