Background: Gold nanoparticles (GNPs) are widely used in nanomedicine as drug carriers, including in targeted radionuclide therapy where therapeutic radionuclides are bound to GNPs. Quantitative assessment of their biodistribution is essential. X-ray fluorescence imaging (XFI) is well suited for detecting high-Z elements, but its quantitative accuracy is compromised by strong attenuation effects, particularly in L-shell XFI where low-energy fluorescence (~10 to 12 keV) is heavily absorbed in tissue. Methods: We developed a computed tomography (CT)-guided attenuation correction algorithm for L-shell XFI. The method generates energy-dependent attenuation maps from co-registered CT data and performs voxel-wise corrections along both excitation and emission paths. The approach was tested on an ex vivo murine tumor sample resected three hours after intratumoral injection of 34.7 μg PEG-modified GNPs. Results: Application of the CT-guided correction substantially improved the relative accuracy of L-shell XFI reconstructions compared to uncorrected data. The corrected distribution maps showed consistent mass recovery across different measurement geometries, demonstrating that the algorithm compensates for the theoretically expected attenuation due to heterogeneous biological tissue. Conclusions: This study provides a proof-of-principle that CT-based attenuation correction enables more reliable and quantitative L-shell XFI of GNPs in biological samples. The approach represents a promising step toward accurate nanoparticle biodistribution assessment in biomedical research, including preclinical studies in targeted radionuclide therapy.