This study addresses the critical need for effective decision-making in medication administration in intensive care. Despite existing methods, limitations persist in predicting the necessity of medication based on patient data. An innovative approach is proposed that leverages laboratory results and vital signs to predict medication administration. The project involved processing and analyzing intensive care patient data, including laboratory results, vital signs, and administered medications. A comprehensive dataset was constructed, and three scenarios were developed, each focusing on a different medication. Random Forest, SVM, and Decision Tree models were applied to predict medication administration. The results demonstrate promising potential for improving treatment effectiveness and patient safety. However, limitations such as limited data availability and the lack of a clear analytical path leading to medication administration were encountered. Despite these challenges, the evaluation of different models’ predictive abilities provides valuable insights. This work contributes a new perspective to medication administration decision-making, paving the way for future research and development in this area.