The　latest　video　compression　standard,　High　Efficiency　Video　Coding　(HEVC),　has　greatly　improved　the　coding　efficiency　compared　to　the　predecessor　H.　264/AVC.　However,　equipped　with　the　quadtree　structure　of　coding　tree　unit　partition　and　other　sophisticated　coding　tools,　HEVC　brings　a　significant　increase　in　the　computational　complexity.　To　address　this　issue,　a　coding　unit　(CU)　decision　method　based　on　fuzzy　support　vector　machine　(SVM)　is　proposed　for　rate-distortion-complexity　(RDC)　optimization,　where　the　process　of　CU　decision　is　formulated　as　a　cascaded　multi-level　classification　task.　The　optimal　feature　set　is　selected　according　to　a　defined　misclassification　cost　and　a　risk　area　is　introduced　for　an　uncertain　classification　output.　To　further　improve　the　RDC　performance,　different　regulation　parameters　in　SVM　are　adopted　and　outliers　in　training　samples　are　eliminated.　Additionally,　the　proposed　CU　decision　method　is　incorporated　into　a　joint　RDC　optimization　framework,　where　the　width　of　risk　area　is　adaptively　adjusted　to　allocate　flexible　computational　complexity　to　different　CUs,　aiming　at　minimizing　computational　complexity　under　a　configurable　constraint　in　terms　of　RD　performance　degradation.　Experimental　results　show　that　the　proposed　approach　can　reduce　58.9%　and　55.3%　computational　complexity　on　average　with　the　values　of　Bjonteggard　delta　peak-signal-to-noise　ratio　as　-0.075　dB　and　-0.085　dB　and　the　values　of　Bjontegaard　delta　bit　rate　as　2.859%　and　2.671%　under　low　delay　P　and　random　access　configurations,　respectively,　which　has　outperformed　the　state-of-the-art　fast　algorithms　based　on　statistical　information　and　machine　learning.