The study compared machine learning models (Random Forest, Artificial Neural Networks, Extreme Gradient Boosting – XGBoost, and Naive Bayes) in the diagnosis of sepsis-associated thrombocytopenia in 1447 sepsis patients from January 2013 to December 2023. Thrombocytopenia occurred in 772 patients (53.4%). The data was split into training and test sets in a ratio of 80:20, with 10-fold cross-validation. XGBoost achieved the best performance with an accuracy of 91.10%, an F1 score of 92.31%, an AUROC of 98.60% in cross-validation, and 97.50% in the test set. Artificial Neural Networks had an accuracy of 90.32% and an F1 score of 91.51%, Random Forest an accuracy of 83.75% and an F1 score of 85.32%. All models had AUROC above 90%. The ten most significant variables were identified using feature importance and SHAP analysis.