(1) : The objective of this study was to determine the optimal post-processing model for dynamic cadmium-zinc-telluride single-photon emission computed tomography (CZT-SPECT). (2) : A total of 235 patients who underwent diagnostic invasive coronary angiography within three months of the SPECT and those who had coronary computed tomography angiography (CCTA) before SPECT (within 3 months) were enrolled in this study. Each SPECT study was processed to obtain global and regional stress myocardial blood flow (sMBF), rest-MBF (rMBF), myocardial flow reserve (MFR) and flow difference (FD) estimates obtained with 1-tissue-compartment (1TCM) and net retention (NR) modes, both with and without attenuation correction. (3) : The use of AC led to significantly higher sMBF, rMBF and DF values obtained by 1TCM compared those values derived by 1TCM with NAC; the lowest values of stress MBF and rest MBF were obtained by 1TCM_NAC. The resting flow, MFR and DF were significantly ( < 0.005) higher in the AC model than in NAC. All quantitative variables were significantly ( < 0.05) higher in NR_NAC than in the 1TC_NAC model. Finally, sMBF, rMBF and FD showed significantly ( < 0.05) higher values by using 1TMC_AC compared to NR_AC. (4) : We suggested that 1-compartment and net retention models correctly reflect coronary microcirculation and can be used for clinical practice for evaluating quantitative myocardial perfusion by dynamic SPECT. Attenuation correction is an important step in post-processing dynamic SPECT data, which increases the consistency and diagnostic accuracy of models.