Intracochlear imaging is of great interest clinically because cochlea is the central organ of hearing. However, intracochlear imaging is technologically challenging due to the cochlea’s small size and encasement in bone. The state-of- the-art imaging techniques are not adequate for high resolution cellular imaging to establish diagnosis without destroying the cochlea. We report in situ imaging of intact mouse cochlea using endogenous two-photon excitation fluorescence (TPEF) as the contrast mechanism. TPEF eliminates the need for exogenous labeling and eradicating the staining-induced artifacts. We used a natural, membranous opening into the cochlea, the round window, as the optical access to reach the organ of Corti, requiring no additional slicing or opening. Our approach provides the maximum non-invasiveness in the imaging process. TPEF exhibits strong contrast allowing deep imaging of mouse cochlea with cellular and even subcellular resolution. Inner hair cell, outer hair cell and supporting cell are clearly identifiable in TPEF images. Distinct morphological differences are observed between healthy and noise-exposed cochleae, allowing detection of specific, noise-induced pathologic changes. The TPEF images taken through the round window are correlated with the whole mount sections, verifying their reliability. Compared with one-photon excitation fluorescence (OPEF) confocal microscope and wide-field transmission microscope images taken under the same magnification and resolution, TPEF images demonstrate clear advantages in terms of sharpness, signal to noise ratio and contrast. These capabilities provide a working foundation for microendoscopy-based clinical diagnostics of sensorineural hearing loss.