This paper presents an application of multiple-agent architecture to beat tracking for musical acoustic signals. Beat tracking is an important initial step in computer understanding of music and is useful in various multimedia applications. Most previous beat-tracking systems dealt with MIDI signals and were not based on multiple-agent architecture. Our system can recognize, in real time, temporal positions of beats in real-world audio signals that contain sounds of various instruments. Our application of multiple-agent architecture enables the system to handle ambiguous situations in interpreting real-world input signals and to examine multiple hypotheses of beat positions in parallel. Even if some agents lose track of beats, other agents will keep the correct hypothesis. These agents have been implemented on different processing elements in a parallel computer. Our experimental results show that the system is robust enough to handle audio signals sampled from commercially distributed compact discs of popular songs.