Collision detection is an important aspect in many real-time simulation environments. Due to its nature of high Computation involved, collision detection can contribute to the bottleneck on the system involving large number of interacting objects. This paper focuses on finding options to efficiently cull away object pairs that are not likely to collide in large-scale dynamic rigid-body simulations involving n-body objects. The main idea is to perform time critical computing concept by manipulation of potential bounding volume techniques. In order to take advantage of a fast collision test and a more accurate result, a hybrid collision culling approach based on sphere - or-Dops was used. Based on initial results, this approach shows a potential adaptation to a massive rigid body simulation.