*Research Assistant at Space and Communications Department, University of Science and Technology, Zewail City. **Coordinator, Space and Communications Engineering, University of Science and Technology, Zewail City. Utilization of Neural Network and the Discrepancy between it and Modeling in Quadcopter attitude Ahmed M.ELDakrory * and Mohammad Tawfik ** Zewail City for Science and Technology Abstract Multi-copter or Multi-rotor is a new field and most of the modeling and control research points are interested to improve the performance of the multi-rotors to do its work with no effort and in best way. Here Quadcopter system was proposed due its wide usage. Quadcopters as it is known as quadrotor has been involved in many applications due to its facilitations and speed. They have been used in Search and rescue, Building Inspection after Earth Quicks, Precision Agriculture, Remote farming, Mapping of building, Transportation, Motivation and mine detection. Modeling of the Quadcopter is the most important part of designing as we see all the characteristics of the system and depend on this observations we determine the controller which sustain all the requirements. We investigate two different methods for modeling first one is modeling which depend on the system of equations describing the system motion and performance, the second method using the neural network using NARX model proposed here[5]. The results will show the difference between those two methods. Introduction Quadcopter is most applicable and safe system that can be used in many wide area. But we must have at the beginning a stable and high performance quadcopter that able to do our mission easily. Quadcopter is depend mainly in the attitude stability as without it we cannot handle the other operations. The main actuator of the quadcopter is the rotors which provide the thrust which generate the required motion. The rotors are directed upward to oppose gravity and they placed in square equally distance formation. The main parameter used to control the quadcopter is the rotational speed so we can controlling the quadcopter attitude, altitude and position. I will distinguish equations of motion of the quadcopter from kinematics to the dynamics of the quadrotors to see all the quadcopter system dynamics and make a simulation for the nonlinear system using the main nonlinear system with reducing any assumption. There is many effort put in modeling the Quadcopter and get the best representation for the quadcopter dynamics [1], [2], [3], [4]. I will compare the result got from those equations with the NARX Neural Network model [5]. Quadcopter Controller for the Attitude and the altitude is the main problem; [2] introduce fuzzy logic controller as a main controller for the attitude and distinguish the performance and the stability of the quadrotor but widely the PID controller has been used and gives good and reliable results for the nonlinear system. Quadcopter Modeling I will introduce the modeling technique used to identify all the system dynamics and for our quadcopter system we will start with kinematics to the dynamics equations to describe all the system behavior.