Objective To investigate the correlations of clinical and pathological characteristics with nonsentinel lymph node (NSLN) metastasis in breast cancer patients with 1 and 2 positive sentinel lymph nodes (SLN), and validate the application value of the NSLN metastasis predictive model of Memorial Sloan-Kettering Cancer Center (MSKCC) in this patients population. Methods Using the BLOU DataBase of breast surgery department in the First Bethune Hospital of Jilin University, the information of 294 patients who accepted sentinel lymph node biopsy and diagnosed with 1 and 2 SLN metastasis between April 2010 and October 2017 was retrospectively analyzed. Correlations of NSLN metastasis with patients’ clinical and pathological data were analyzed by Chi-square test and logistic regression analysis. The predictive value of MSKCC nomogram was evaluated by the area under the receiver operating characteristic curve (AUC). Results NSLN metastasis was found in 86 of the 294 patients. The correlations of NSLN metastasis with lymphovascular invasion and SLN positive number/total SLN number were statistically significant (P < 0.05). The AUC value of the MSKCC predictive model for NSLN metastasis in this study population was 0.643. Conclusions 1 and 2 SLN positive patients with high SLN metastasis rate or with lymphovascular invasion are more likely to have NSLN metastasis. The accuracy of MSKCC nomogram in predicting NSLN metastasis is low for breast cancer patients in the First Bethune Hospital of Jilin University.