The sensory epithelia of the inner ear contain mechanosensitive hair cells that transmit sound, gravity and head motion signals. This protocol describes an in vitro 3D differentiation method, by which the inner ear sensory epithelium harboring hair cells are derived from human pluripotent stem cells (hPSCs). To begin the differentiation, hPSCs are aggregated in low-binding 96-well plates and treated with extracellular matrix proteins to promote epithelialization. By recapitulating signaling pathway activation and attenuation during in vivo inner ear development, the aggregates are treated with small molecules and recombinant proteins that modulate signaling pathways such as BMP, FGF and WNT in a stepwise manner. These treatments induce sequential formation of non-neural ectoderm (NNE), otic-epibranchial progenitor domain (OEPD), and otic placodes. The otic placodes subsequently undergo self-guided morphogenesis to form otic vesicles, which eventually give rise to sensory epithelia containing inner ear hair cells and supporting cells, as well as neurons forming synapses with the hair cells. These hPSC-derived inner ear sensory structures are designated human inner ear organoids. As human inner ear biopsies are nearly impossible to obtain without causing severe injuries to the auditory system of the patients, the human inner ear organoid system provides a powerful in vitro platform for studying human inner ear disease and development.