The storage and seepage space of shale is mainly composed of pores and fractures with complicated distribution patterns. The calculation accuracy of pore structure parameters largely affects the reservoir evaluation results. Taking the Niutitang Formation in the Cengong area of Guizhou as an example, we have performed pore structure evaluation. First, with the Archie formula, the process of wellbore mud intrusion is approximated as the process of laboratory high-pressure mercury intrusion. Based on conventional and nuclear magnetic resonance (NMR) logging data, a new model for the conversion from spectrum to pseudocapillary pressure curve is deduced. Then, the model is calibrated with the data of the high-pressure mercury intrusion test and the pore structure parameters such as reservoir pore-size distribution curve and maximum pore throat radius are calculated with the calibrated model. The results show that the maximum pore throat radius and total porosity calculated with NMR logging data are relatively reliable. The calculated median radius has the medium error, whereas the displacement pressure and median pressure have the large error. The pore volume percentage of pores with a radium of 1–10 μm is up to 60%, and the well-developed microfractures are beneficial to the fracturing of the reservoir. Therefore, NMR logging data combined with conventional logging data can better reflect the pore structure characteristics of reservoirs. The study provides a basis for the identification of complex reservoirs and the qualitative prediction of productivity.