The Tarim Oilfield has multiple oil-bearing formations and abundant oil and gas resources. As the oilfield exploration further develops, the geological conditions of the exploration targets have become more complex, with diverse lithology, complex pore structure, and strong reservoir anisotropy. Thin reservoirs have gradually become the main exploration targets. Lithological density logging is an important method for determining formation lithology, calculating porosity, and identifying oil and gas reservoirs. The existing three detector density logging tool compensates by placing a backscatter detector within the negative source distance range, which improves the vertical resolution and measurement accuracy of the instrument to a certain extent. However, due to the large wellbore size and high mud density, it is still difficult to meet the logging evaluation requirements of thin reservoirs. By analyzing the relationship between photon flux and source distance, formation density, different wellbore sizes, and fluid density inside the wellbore, a calibration model for relevant logging data is determined, and the selection of high, medium, and low energy windows is optimized. On the basis of the above research, optimize the data processing method of the whole energy spectroscopy, use the formation information contained in the spectrum and the measurement information of multiple detectors to calculate the density and photoelectric absorption index of the formation, and further improve the vertical resolution and measurement accuracy of the three detector density logging tool.