Quantitative measurement of articular cartilage using optical coherence tomography (OCT) is a potential approach for diagnosing the early degeneration of cartilage and assessing the quality of its repair. However, a non-perpendicular angle of the incident optical beam with respect to the tissue surface may cause uncertainty to the quantitative analysis, and therefore, significantly affect the reliability of measurement. This non-perpendicularity was systematically investigated in the current study using bovine articular cartilage with and without mechanical degradation. Ten fresh osteochondral disks were quantitatively measured before and after artificially induced surface degradation by mechanical grinding. The following quantitative OCT parameters were determined with a precise control of the surface inclination up to an angle of 10° using a step of 2°: optical reflection coefficient (ORC), variation of surface reflection (VSR) along the surface profile, optical roughness index (ORI) and optical backscattering (OBS). It was found that non-perpendicularity caused systematic changes to all of the parameters. ORC was the most sensitive and OBS the most insensitive to the inclination angle. At the optimal perpendicular angle, all parameters could detect significant changes after surface degradation (p < 0.01), except OBS (p > 0.05). Nonsignificant change of OBS after surface degradation was expected since OBS reflected properties of the internal cartilage tissue and was not affected by the superficial mechanical degradation. As a conclusion, quantitative OCT parameters are diagnostically potential for characterizing the cartilage degeneration. However, efforts through a better controlled operation or corrections based on computational compensation mechanism should be made to minimize the effects of non-perpendicularity of the incident optical beam when clinical use of quantitative OCT is considered for assessing the articular cartilage.