Background and Aims Venous thromboembolism (VTE) is an important complication after spine surgery. Although the Caprini score is widely used for perioperative VTE risk stratification, it has limited ability to capture early postoperative information related to coagulation and fibrinolysis, nutritional status, and surgical burden. This study aimed to evaluate the incremental predictive value of early postoperative clinical information beyond the postoperative Caprini score and to develop and internally validate models for predicting in-hospital VTE after spine surgery.Methods This single-center retrospective cohort study included 1 508 adult patients who underwent spine surgery at Xiangya Hospital, Central South University, in 2024. The primary outcome was overall VTE occurring during postoperative hospitalization. Postoperative laboratory variables were defined as the first measurements obtained within 24-48 h after surgery; for patients who developed VTE, only measurements obtained before VTE diagnosis were used. Three Logistic regression models were developed. Model A included the postoperative Caprini score alone. Model B additionally included age, albumin, fibrin degradation products (FDP), operation duration, intraoperative blood loss, and transfused red blood cell units. Model C included the postoperative Caprini score, age, albumin, and FDP. Internal validation was performed using 1 000 bootstrap resamples and stratified 10-fold cross-validation. Model discrimination, calibration, and clinical utility were assessed using the area under the receiver operating characteristic curve (AUC), calibration analysis, and decision curve analysis. A preoperative Caprini score model, a sensitivity analysis excluding FDP, and an auxiliary LASSO analysis were also performed. Based on the validated models B and C, online web-based risk assessment tools were developed to enable individualized VTE risk calculation and visual assessment.Results Among the 1 508 patients, 156 developed VTE, corresponding to an incidence of 10.34%. After bootstrap optimism correction, the AUC of models A, B, and C were 0.706, 0.824, and 0.821, respectively. The corresponding out-of-fold AUC from 10-fold cross-validation were 0.699 (95% CI=0.656-0.742), 0.824 (95% CI=0.792-0.855), and 0.820 (95% CI=0.790-0.851). Models B and C demonstrated significantly better discrimination than Model A (both P<0.001), whereas no significant difference was observed between models B and C (P=0.494). The calibration slopes of models B and C were 0.954 and 0.975, respectively. Decision curve analysis showed that models B and C generally provided greater net benefit than model A across threshold probabilities of approximately 0.02-0.30. After excluding FDP, the out-of-fold AUC of the enhanced model was 0.795; paired DeLong comparison with model B yielded an AUC difference of 0.028 (P=0.006), indicating an incremental predictive contribution of FDP. Online web-based risk assessment tools were developed based on the validated models B and C, enabling individualized VTE risk calculation and visual assessment.Conclusion The postoperative Caprini score provides baseline risk information for in-hospital VTE after spine surgery but has limited discrimination when used alone. Incorporating age, albumin, FDP, and surgery-related information further improves risk discrimination. The simplified model maintained predictive performance comparable to that of the full enhanced model while using fewer predictors, supporting its potential use for postoperative VTE risk stratification and online assessment. External validation in independent multicenter cohorts is required before broader clinical application.