To　improve　the　convergence　rate　of　path-following　errors　of　a　bionic　snake　robot　and　reduce　the　overshoot　of　kinematic　errors,　a　path-following　control　method　based　on　fuzzy　line-of-sight　(LOS)　guidance　for　a　bionic　snake　robot　with　unknowns　is　developed.　This　method　designs　the　time-varying　LOS　forward　distance　and　sliding　mode　switching　gain　according　to　fuzzy　rules,　offsets　the　external　interferences,　weakens　the　damage　of　chattering　to　robot　joints,　and　reduces　the　negative　impact　of　external　interferences　on　the　robot　path-following.　In　addition,　the　uncertain　friction　coefficients　are　predicted　by　using　the　adaptive　estimation,　which　compensates　for　joints＇　rotating　force　inputs　and　improves　environmental　adaptability　of　the　body.　By　establishing　the　appropriate　Lyapunov　function,　the　stability　of　the　gait　angle,　motion　position,　and　whole　system　are　verified.　Simulation　and　experiment　results　show　the　feasibility　and　superiority　of　the　proposed　method.　©　1996-2012　IEEE.