This article presents an overview of emotional recognition of music (MER) through the fuzzy logic approach. First, related papers are reviewed to analyze how fuzzy logic is used in each phase of a typical MER process. Subsequently, a prototype of a fuzzy system is designed to classify musical pieces by arousal levels, defining the tempo of the songs as the system's input and the level of arousal as the output. Based on the review of the literature and the results obtained with the fuzzy system prototype, a discussion is presented to improve the understanding of the main differences between classification systems with fuzzy logic or with machine learning approaches, focusing on the success rate of MER, especially in the labeling and classification processes. This review and the comparison of both approaches (fuzzy logic and machine learning) reveal fuzzy logic's principal contributions, expand the knowledge of the current developments in MER and present possible improvements for the design of MER systems.