Conformal coating is a protective coating widely used for protecting printed circuit board (PCB) from harsh environmental conditions. The conformal coating’s thickness is one of the key parameters for evaluating the quality of conformal coating, and thus the thickness needs to be measured before PCB assembling. In this paper, we proposed to utilize spectral domain optical coherence tomography (SD-OCT) combined with an image segmentation algorithm to achieve a rapid and non-destructive thickness measurement on conformal coating. The spectral domain OCT system was built with a broadband SLD light source (bandwidth: 180 nm) so that an axial resolution of 1.72 μm was obtained. Meanwhile, an image segmentation algorithm based on boundary tracing was designed to segment the conformal coating in OCT image for rapid and accurate thickness measurement. To evaluate the proposed method, the measurement results in this paper were compared with the traditional metallographic examination, analyzing the consistency of the detected boundaries as well as the measured thickness between these two methods. In addition, we also compared the method proposed in this paper with the edge detection algorithm based on image gradient previously reported by our group in terms of the accuracy and efficiency, so as to justify the advantages of this newly developed method. The results show that the measurement method designed in this paper is in good agreement with the traditional metallographic method, and the thickness of the coating can be accurately measured. Based on the three-dimensional imaging capability of the system, the thickness topographic map of the conformal coating can be visualized clearly, which cannot be realized by the metallographic section method. Compared with the method proposed by our group previously, this method can provide more accurate and efficient measurement, paving the way for real-time and inline thickness measurement on conformal coating.