This study aimed to examine the association between the maximum standardized uptake value (SUVmax) of different molecular subtypes of primary breast cancer with axillary lymph node (ALN) metastasis. Methods: The medical records of 633 patients, who underwent 18F-fluorodeoxyglucose positron emission tomography-computed tomography (PET/CT) for preoperative staging, were retrospectively reviewed. The cohort was stratified by the following molecular subtypes with immunohistochemical examination: luminal A, luminal B, human epidermal growth factor receptor 2 positive (HER2+), and triple negative. We evaluated the optimal cutoff SUVmax to predict ALN metastasis in each subtype using the receiver operating characteristic (ROC) analysis. Moreover, the risk factors for ALN metastasis were evaluated. Results: Overall, the SUVmax was positively correlated with the number of metastatic ALN (p= 0.001). The mean SUVmax was higher in aggressive subtypes (4.5± 0.2, 6.1± 0.4, 6.5± 0.5, and 7.5± 0.5 in luminal A, luminal B, HER2+, and triple negative, respectively, p< 0.001). Upon ROC analysis, the SUVmax of the HER2+ subtype predicted ALN metastasis most accurately, with a cutoff value of 5.5, area under the curve (AUC) of 0.708, sensitivity of 74.2%, and specificity of 64.6% (p= 0.002). The triple negative subtype did not show a significant difference in SUVmax between patients with and without metastasis (p= 0.13). Subtype-adjusted SUVmax, HER2 positivity, lymphovascular invasion, and T stage were significant predictors for ALN metastasis. Conclusion: The SUVmax of primary breast cancer may be an independent predictor of ALN metastasis, being the most accurate in the HER2+ subtype. As PET/CT could facilitate tailored axillary management, this approach could be considered for the initial staging and treatment planning in patients with breast cancer.