When being pursued by guided munitions, a fixed wing aircraft is likely to attempt to avoid interception. If a team of autonomous missiles can learn how their motion affects the induced motion of their target, the exploitation of this knowledge can facilitate controlled diversion and interception of the target. Motivated by recent advances in the field of herding control, this paper details a novel control and estimation strategy for a team of missiles tasked with diverting a target aircraft from its planned path and intercepting it somewhere on a prescribed “safe" trajectory. A neural network-based estimation scheme is used to approximate the uncertain missile-target interactions online. The missile controllers leverage these estimates to ensure that the diversion and interception objectives are achieved. A rigorous Lyapunov-based analysis examines the stability of the closed loop error system.