Quadcopter is a type of UAV having two pairs of counter rotating rotors. Its movement isrncontrolled by adjusting the relative speed or relative trust and torque of each rotor which arernspun by electric motors. However, quadcopter is nonlinear, underactuated MIMO systemrnand also its inputs and outputs are constrained, which provide a suitable platform for controlrnalgorithm development and investigation of both stabilization and trajectory trackingrncontrol.rnBecause quadcopter is severely under-actuated and coupled system, a cascade control approachrnis proposed and applied for trajectory tracking control. Three di erent MPC methods,rnnamely Linear MPC, Feedback linearization based MPC and Nonlinear MPC are designedrnfor attitude control. Moreover, the control of x, y, and z position is also addressed,rnby utilizing a conventional PID controller, to test the performance of the attitude controllersrnduring trajectory tracking control.rnThe nonlinear mathematical model of a quadcopter's dynamics, solved from Newton's andrnEuler's laws, is investigated by realizing the designed controllers using MATLAB/Simulink.rnThe introduced attitude controllers are compared based on three performance evaluationrnfactors, tracking accuracy, control e ort e ciency and output disturbance rejection capacity.rnThe performance of the proposed control schemes are veri ed by comparative simulationrnresults and Root Mean Square Error (RMSE) tracking accuracy performance measure. Thernsimulation results shows Linear MPC o er poor performance even if it is simple to design andrno er fast response compared with the other. Feedback linearization based MPC ensure thernutilization of linear MPC for nonlinear plant, in this case for nonlinear quadcopter model. Inrnaddition, Feedback linearization based MPC and Linear MPC strategies couldn't withstandrnoutput disturbances. Moreover, Feedback linearization based MPC strategy is incapable torndeal with input constraints. Conversely, nonlinear MPC o ers a good performance withrnconstraints and output uncertainty. Even if it is computationally intensive.