Abstarct: One of the main challenges in using unmanned aerial vehicles (UAVs) as communication relays and fly baxse stations is finding the optimal altitude and position for static and dynamic mobile users. Failure to fine-tune the altitude and position of the UAV can reduce the average coverage probability, increase the distance between the user and the UAV, increase the average path loss, and reduce the average achievable rate for Downlixnk users. In this dissertation, we present a network with two different scenarios. The first scenario includes a fixed-position drone, Downlixnk users (DUs) with a fixed position, and the Device-to-Device (D2D) users with a fixed-position. Another scenario includes, a limited-speed Moving UAV, Downlixnk users with a random Gauss-Markov mobility model and D2D users with a fixed position. First, with a completely new approach baxsed on the golden section search (GSS) method and considering the UAVs Transmit Power, we calculate the optimal altitude of the UAV, which is the minimum function of average path loss, with the least processing time compared to other methods for the first and second scenarios. Then, to find the optimal position of the drone in the second scenario and baxsed on the Global Positioning System (GPS) in a non-real time (Non-Real Time) mode, by considering several random points in the desired area, measure the distance to random points with Using two methods and finally searching using an heuristic algorithm, we get the optimum position (optimal random point) of the drone for all time slots. The simulation results using MATLAB software show that after finding the optimum position and altitude, the average coverage probability is 0.95 or the total number of users served 95 out of 100 users, as well as the average achievable rate for the Downlixnk users, 0.15 Giga Bit per seconds (Gbps) for 50 slots is calculated when it has the maximum value compared to other methods. The processing time to find the optimal altitude value using the golden section search algorithm is estimated to be 0.03 seconds.