Pore structure is a key parameter used to evaluate reservoir quality.At present, experimental method is the most important method to analyze reservoir pore structure.However,coring data may be limited, and it is not possible to perform experimental analyses on all cores.Therefore,the researchers explored the use of logging techniques to study the pore structure of reservoirs.The relationship between pore geometry index (PGI),pore permeability and mercury injection parameters was analyzed based on mercury injection experiment,thin slice analysis,production test data and well logging data.These results can then determine the response characteristics of the logging parameters that correspond to different pore structures and establish a method of modeling the PGI through a multiple parameter regression and neural network method.This research shows that:(1)PGI can quantitatively characterize pore structure, and the maximum pore throat radius,displacement pressure and flow unit index have the highest correlation with PGI,which can be accurately characterized by practical formulas;(2)Natural gamma ray,natural potential amplitude difference, acoustic transit time,density,compensated neutron,deep and shallow resistivity logging data can reflect the quality of the reservoir pore structure.However,there are limitations in evaluating reservoir pore structure with a single logging parameter.Multi-parameter regression method and neural network method realize the quantitative calculation of pore structure from the perspective of multi-parameter and nonlinear.(3)The neural network method and multiple parameter regression method are used to study the pore structure of reservoir and realize the continuous quantitative calculation of pore structure index in a single well.It can be used in uncored analysis Wells and as one of the parameters to evaluate reservoir pore structure.