Autism Spectrum Disorder (ASD) is characterised by deficits in emotion recognition (ER) and emotion awareness, impacting considerably their ability to develop social relationships. ER deficits were also thought to be the reason for the apparent lack of responsiveness to the emotional component of music in ASD. However, findings from previous research indicated that individuals with ASD and neurotypicals did not differ in terms of physiological arousal due to music-evoked emotions (Heaton et al. 2008; Allen et al. 2012) and deficits in ER could be due to a reduced ability to verbally express emotions and recognise them in themselves, mediated also by verbal abilities (Bird et al. 2010; Cook et al. 2013). A music-based intervention was designed to train ER and emotion awareness in adolescents with ASD. 11 adolescents (1 female) with a previous ASD diagnosis, took part in the intervention that involved 5 consecutive music sessions and pre and post-testing sessions. During the intervention, emotionally classified musical excerpts were presented, which triggered a posterior discussion about how emotions are displayed in faces, voices and daily experiences. Results showed a significant difference in pre and post-intervention scores in ER in voices, and significant correlations between verbal abilities and ER scores. According to previous research relating the level of Alexithymia and ER abilities, most verbally able participants achieve significant improvements after the intervention. These results suggest the potential benefit of music-based interventions to enhance ER in adolescents with ASD and depicted a relationship between verbal abilities and ER skills in ASD.