A quadrotor is a rotary-wing Unmanned Aerial Vehicle (UAV) that has four rotors. The quadrotor control and autonomy is a challenging task as the quadrotor system is considered underactuated system. The aim of this paper is to present the specification and the implementation of an autonomous quadrotor delivery system that fuses data from camera and IMU (Inertial Measurement Unit) to allow for high precision visual-inertial navigation. In this paper, an autonomous quadrotor delivery system is proposed. The specification of the system is proposed. The implementation of the system introduced SolidWorks design for the system, mathematical modelling for the pickup and delivery problem (PDP). Simulation results using MATLB have been shown for the implementation of the model, control system and the simultaneous localization and mapping (SLAM) using Extended Kalman Filter algorithm (EKF-SLAM) used for making the system autonomous. Experimental results for the use of canny edge detection algorithm have been shown.