Micro Air Vehicles (MAV) can deliver their full potential only if they are equipped with autonomous navigation capabilities. Optic flow has emerged as a significant tool for many autonomous navigation tasks. Optic flow fields remain to some extent uncertain because of image noise, lighting changes, low contrast regions, multiple motions in single localized regions and so on. Tracking multiple targets by itself is a challenging problem; it is made more challenging by three factors in the MAV scenario: the uncertainties in optic flow, 6 degrees of freedom (DoF) of movement of the platform and possible 6 DoF of movement of the targets. Finite set statistics provides a unified probabilistic foundation for multisource-multitarget tracking; its computational complexity has been addressed by Probability Hypothesis Density filters and many approximation techniques. We propose a statistical model for tracking multiple objects (targets) from optic flow information obtained from nose-mounted camera. Significant contributions of this paper are enhanced state-transition and measurement models and use of random set theory to track varying number of targets. A lightweight mechanism for data association is also proposed.