The emergence of mobile edge computing (MEC) technology provides computing support for resource-intensive and delay-sensitive applications. However, fixed locations and high deployment costs limit the development of MEC. The existing MEC technology is not applicable when the network facilities are sparse, and the number of multimedia users increases sharply. The combination of low-cost and highly mobile drones and MEC can provide high-quality services for wireless communication. This paper proposes a novel MEC system assisted by a single UAV, in which multiple ground multimedia devices are served by a UAV equipped with a server. Computing tasks on multimedia devices can be done locally or offloaded to a UAV equipped with a server. The UAV is highly fixed in the air, providing computing services in an orthogonal multiple-access mode within the time frame. The offloading strategy is optimized to maximize the task calculation rate in each time frame subject to the constraints of discrete binary, computational resource, energy consumption, and maximum delay. This optimization problem is non-convex. We propose a Deep Deterministic Policy Gradient (DDPG) algorithm to solve this problem. Compared with the baseline model, the algorithm can maximize the system computing rate and effectively save computing resources in the wireless system of single UAV-assisted edge computing.