<p>The deviation degree of the fuzzy number is an essential factor in portraying fuzzy characteristics. Two different fuzzy ranking methods that take deviations into account have been proposed by Wang et al. and Asady, which have received a lot of attention, but have since been shown to have various counterexamples. To solve this problem, this study first introduces these two ranking methods systematically and outlines their shortcomings. Secondly, we propose a novel hybrid two-dimensional deviation degree to rank L–R fuzzy numbers based on the two-dimensional centroid. In addition, three comparison examples demonstrate that the hybrid two-dimensional deviation degree-based ranking method can effectively overcome the shortcomings of the existing methods in comparing L–R fuzzy numbers, such as logic errors, the violation of intuition, poor discrimination, etc. Finally, the transferability and rationality required for the hybrid two-dimensional deviation degree ranking method are discussed.</p>