Logarithmic ratio of positive lymph nodes (LODDS), number of positive lymph nodes (NPLN), and number of lymph nodes to positive lymph nodes (pLNR) are three lymph node classifications; however, their function in prognosis is unclear. To establish and validate an optimal nomogram according to the comparison among the 7 TNM stage of American Joint Committee on Cancer (AJCC) and the three lymph node classifications. A total of 881 patients from the Surveillance, Epidemiology and End Result (SEER database) with TNM in laryngeal squamous cell carcinoma from 2000 to 2018 were involved. The enrolled patients were allocated randomly into a training cohort and a validation cohort. Univariate cox regression analysis and multivariable cox regression analysis were applied to explore the predictors. The Akaike Information Criterion (AIC) and Harrell's concordance index (C-index) were to measure the predictive value and the accuracy of the prognostic models. Moreover, integrated discrimination improvement (IDI) and net reclassification index (NRI) were also used to assess the predictive abilities to models. According to the optimal model, nomograms were established and compared with 7 TNM stage of AJCC via the decision curve analysis. NPLN, LODDS, and pLNR were three predictors for the overall and cancer-specific survival in the larynx squamous cell carcinoma. According to the AIC, C-index, IDI, and NRI, the model of NPLN combined with LODDS was assumed as the optimal prognostic model. Moreover, the decision curve analysis suggested that the nomogram demonstrated a better predictive performance, compared with the 7 AJCC TNM stage. The proposed nomograms we constructed for larynx squamous cell carcinoma has potential in the prediction of patients after surgery.