In recent years, the mental workload of knowledge workers has received heightened attention, driven by its significant influence on productivity and its critical implications for overall well-being and mental health. Knowledge workers often face high mental demands, particularly in planning and coordination tasks, leading to stress and reduced efficiency. While several machine learning (ML) models have been employed to predict mental workload, their accuracy has remained below optimal. This study introduces an ensemble classification model that combines K-Nearest Neighbors (KNN), Support Vector Machine (SVM), and Random Forest (RF) through a voting ensemble algorithm to classify the mental workload of knowledge workers. Utilizing the SWELL-KW dataset, including physiological and subjective data, the proposed model achieved a 97% accuracy rate, outperforming individual ML models. These findings indicate that the ensemble model offers a promising approach to enhancing the prediction of mental workload, providing companies with a powerful tool to address mental health challenges in the workplace better.