To better understand the limitations of high-resolution adaptive optics scanning laser ophthalmoscopy (AOSLO), we describe an imaging model that examines the smallest cone photoreceptors in the fovea of normal human subjects and analyze how different factors contribute to their resolution. The model includes basic optical factors such as wavelength and pupil size, and defines limits caused by source coherence which are specific to the AOSLO imaging modality as well as foveal cone structure. The details of the model, its implications for imaging, and potential techniques to circumvent the limitations are discussed in this paper.