Gears and gear systems, like any other mechanical system, are subjected to design parameter, and loading uncertainties emanating from inherent randomness, manufacturing, and assembly errors. The traditional deterministic approach to the design of such systems overlooks these uncertainties. This work presents a novel probabilistic multibody dynamic analysis (PMBDA) that enhances the deterministic design practice of gears and gear systems. A contact based, rigid multibody spur gear pair model with random loading, and design parameters has been developed. An advanced mean based on fast probability integration method was implemented to perform a reliability analysis of performance measurements: dynamic factor, root bending stress, and fatigue life of gears. Probabilistic sensitivity analysis of these performance functions to several random variables was also determined. In addition to revealing system reliability or probability of failure, the PMBDA approach also helps designers to consider certain variables critically.