A talking and singing robot which has equivalent mechanical organs to human vocal system is being developed based on a mechatronics technology under a feedback control. While various ways of vocal sound production have been actively studied so far, a mechanical construction of the vocal system is considered to advantageously realize natural vocalization with its fluid dynamics. Motors are employed for the manipulation of the mechanical system. The robot adaptively learns the relations between motor control parameters and the generated vocal sounds using an auditory feedback learning with neural networks, and sings a song by mimicking a human vocalization. This paper presents the construction of the talking robot and its singing performance, together with the adaptive control for the pitch and phoneme learning. The robot generates vowel and consonant sounds of different pitches by dynamically controlling the vocal cords, vocal tract and nasal cavity.