This research paper explores the classification of turkish music based on emotional content using machine learning techniques. Utilizing a dataset comprising 400 music samples evenly distributed across four emotional categories—happy, sad, angry, and relax—each sample is represented by 50 acoustic features derived from diverse genres of Turkish music. Initial experiments with individual machine learning models demonstrated promising results with accuracy of 92.5% by Gradient Boost Classifier. To enhance classification performance, this study utilizes multi-modal classifiers through stacking and voting methods, leveraging the strengths of various models. Stacking and Voting classifiers achieved an accuracy of 93% on the dataset which indicate that multi-model classifiers can achieve superior accuracy and robustness compared to individual models, offering significant implications for applications in personalized music recommendation systems, therapeutic tools, and interactive entertainment systems. This paper comprehensively evaluates individual and multi-model classifiers, detailing their methodologies, experimental results, and potential for future advancements in music emotion recognition.