This study proposes a new approach for the classification of the calls detected in the songs with the use of Hidden Markov Models (HMMs) based on the concept of subunits as building blocks. HMMs have been used once before for such task but in an unsupervised algorithm with promising results, and they are used extensively in speech recognition and in few bioacoustics studies. Their flexibility suggests that they may be suitable for the analysis of the varied repertoire of humpback whale (Megaptera novaeangliae) calls because they cope well with variations in the call durations, which is a common feature in humpback whale vocalizations. Another attractive characteristic of HMMs is that highly developed tool-set is widely available as a consequence of the widespread use of their employment for human speech analysis.We describe the HMM classification method and show that a high level of performance can be achieved with modest requirements both in terms of computational load and storage. Training stage requires minimal manual input and once trained the recognition process is fully automated. We will present how the classification performance is affected by different amount of training.