This paper presents music learning theory of Gordon and shows how it can be applied to music classes. The music learning theory of Gordon emphasizes on learning audiation capability. Audiation is hearing in imagination music that is not played in reality Audiation capability is essential to achieve the purpose of music education, that is, the understanding and enjoyment of music. Gordon insists that to develop audiation capability students be taught patterns which are basic words of music. When teaching patterns, tonal patterns and rhythm patterns should be taught independently of each other, that is, tonal patterns without rhythm aspect and rhythm patterns without tonal aspect. Moreover, Gordon claims that music should be taught in sequence by following eight levels of learning. The eight levels can be grouped into discrimination learning and inference learning. Discrimination learning is rote learning from teachers and inference learning is self-learning by which students derive new knowledge from familiar ones. Discrimination learning consists of aural/oral, verbal association, partial synthesis, symbolic association, composite synthesis. Inference learning consists of generalization, creativity/improvisation, theoretical understanding. The movement from the current level to the next should happen only when the current level of learning is sufficiently achieved. Each music class consists of Learning Sequence Activities and Classroom Activities. The former refers to teaching tonal patterns and rhythm patterns and the latter refers to singing, listening, and performing. Learning Sequence Activities develop skills needed for Classroom Activities. The details of both activities are different depending on the levels in learning sequence in which they happen. Finally this paper presents a teaching plan created by applying Gordon`s learning theory to the aural/oral learning level.