Optical coherence tomography (OCT) is a non-invasive method to quantify neurodegeneration as an outcome in multiple sclerosis clinical trials; however, no data exist on Cirrus spectral domain optical coherence tomography (SD-OCT) reproducibility in patients with multiple sclerosis. The objective of this study was to determine the protocol for achieving optimal inter-visit, inter-rater, and intra-rater reproducibility for studies performed on healthy controls and multiple sclerosis patients utilizing novel high-definition SD-OCT. This is a prospective study of inter-visit, inter-rater, and intra-rater reproducibility in multiple sclerosis patients (n = 58) and healthy controls (n = 32) on Cirrus-HD SD-OCT. Excellent reproducibility of average and quadrantic retinal nerve fiber layer (RNFL) thickness values, average macular thickness (AMT), and total macular volume (TMV) [measured by intraclass correlation coefficient (ICC)] was found for inter-visit (healthy controls: mean RNFL = 0.97, quadrant range = 0.92-0.97, AMT = 0.97, TMV = 0.92), inter-rater (MS: mean RNFL = 0.97, quadrant = 0.94-0.98, AMT = 0.99, TMV = 0.96; healthy controls: mean RNFL = 0.97, quadrant = 0.94-0.97, AMT = 0.98, TMV = 0.99), and intra-rater (MS patients: mean RNFL = 0.99, quadrant = 0.83-0.99, AMT = 0.97, TMV = 0.98) reproducibility. The reproducibility of retinal measures derived by Cirrus HD-OCT, especially quadrantic values, is excellent. Specific procedures for OCT acquisition and analysis of retinal imaging metrics using SD-OCT technology may improve the application of this novel technology in multiple sclerosis.