This work investigates the collision avoidance visual coverage control problem of multiple second-order Mobile Aerial Agents (MAAs). When performing coverage control tasks, MAAs need to consider both the sensing model and the sensing quality. An integrated coverage performance function is defined to evaluate the coverage effect. For the coverage control problem, a novel distributed control law is proposed to drive the second-order MAA system to converge to a locally optimal coverage configuration, and the stability of the corresponding coverage control system is proven by introducing a Lyapunov-like function. Moreover, a potential function term is employed to avoid collision between the MAAs and improve the safety performance of the system. Finally, numerical examples are given which demonstrate the effectiveness and practicality of the proposed control law.