The survival rate varies tremendously in T1a N0 M0 glottic cancer patients, which may be associated with the difference on patients' characters, such as age, treatment, and marital status. The objective of this study is to identify the vital factors and construct a dynamic prognostic model for predicting the cancer-specific survival (CSS) of patients with T1a N0 M0 glottic cancer.The data of T1a N0 M0 glottic cancer patients were extracted retrospectively from the SEER database between 2004 and 2015, which were randomly divided into the training dataset (70%) and the validation dataset (30%). The training cohort was used to identify independent prognostic factors and build the prognostic nomogram. While the validation cohort was used to validate the applicability of the newly constructed model.The model was evaluated with the discrimination, the calibration, and the clinical benefit. The external data collecting from a medical center were used to validate the performance of the prognostic model.Totally, 3286 eligible patients in the data of the SEER database and 139 eligible patients in our external clinical cohort were finally identified. 5 independent prognostic factors (age, marital status, Grade, primary site surgery, and chemotherapy) were identified and applied to develop the dynamic prognostic tool. C-indexes, receiver operating characteristic curves, calibration curves, and decision curve analyses proved that the prognostic nomogram showed excellent predictive accuracy, ability, and prognostic value. Using internal and external cohorts, the validation of the model also proved its reliability.Prognostic factors for T1a N0 M0 glottic cancer were identified. The novel web-based prognostic prediction tool may be beneficial in clinical decision-making and has value in risk stratification, personalized treatment, and clinical trial design.© 2023 John Wiley & Sons Ltd.