During　the　flight　of　quadrotor　unmanned　aerial　vehicle　manipulator,　external　wind　and　model　uncertainties　will　significantly　affect　the　accuracy　and　stability　of　the　controller.　This　study　investigates　the　problem　of　high-precision　attitude　control　for　the　quadrotor　manipulator　that　is　equipped　with　a　2-DOF　robotic　arm　in　presence　of　several　disturbances　based　on　a　new　sliding　mode　observer　and　a　corresponding　sliding　mode　controller.　As　for　the　proposed　sliding　mode　observer,　to　reduce　its　estimation　chattering　phenomenon,　a　sigmoid　function　is　exploited　to　replace　the　discontinuous　signum　function.　Moreover,　to　adapt　to　the　estimation　of　disturbances　with　different　upper　bounds　of　the　first　derivative,　a　fuzzy　logic　system　algorithm　is　used　to　adaptively　update　the　observer　gains　and　slope　parameters　of　the　sigmoid　function.　Furthermore,　a　generalized　super-twisting　algorithm　is　incorporated　into　the　proposed　sliding　mode　observer.　Similarly,　the　sliding　mode　controller　is　constructed　by　using　the　generalized　super-twisting　algorithm　and　a　sigmoid　function　in　addition　to　the　sliding　mode　observer-based　feedforward　disturbance　compensation.　In　addition,　to　further　relieve　the　influence　of　system　sensor　noise,　a　tracking　differentiator　is　exploited　to　incorporate　with　both　proposed　sliding　mode　observer　and　sliding　mode　controller.　Finally,　to　demonstrate　the　effectiveness　of　the　proposed　method,　several　simulations　and　experiments　on　the　quadrotor　unmanned　aerial　vehicle　manipulator　system　are　conducted　under　varying　external　disturbances.