Both CL-TSMC and CL-NTSMC are proposed to control the dynamics of quadrotor in presence of unknown dynamics,time-varying disturbances and input dead zone. The SLFN is employed to approximate system uncertainty while the DOB is constructed to deal with the compound disturbance. The finite-time convergence is guaranteed via Lyapunov analysis. Simulation results show that the composite estimator can follow uncertainty and the proposed method can efficientlyimprove the tracking performance of quadrotor.