Most discrete event simulation (DES) of equipment failure and reliability modeling in mining uses global reliability data i.e., time between failure (TBF) and time to repair (TTR) of material handling equipment to simulate failure and its effect on production losses. However, raw data from most mines contains data on failure of different components of the equipment. It would appear that, to accurately incorporate equipment failure and reliability into simulation models, the failure data should be classified into sub-system data and then used in the model to simulate failure and repairs. Recently, many authors (Awuah-Offei et al. 2003; Yuriy and Vagenas 2008) have used DES as a modeling tool to model material transfer systems in mining. However, equipment breakdowns and corresponding maintenance processes were included in the model up-to the machine level (globally). The objective of this work is to examine the effect of including or not including subsystem reliability data on model performance. Reliability and maintenance data were collected, sorted, and analyzed from a hard rock auriferous surface mine. A discrete event simulation of truck-shovel material transfer was built in Arena® and validated with the field data. The model was used to evaluate the effect of simulating failure at the subsystem or global level on total production. Simulation result show 4% reduction in total production when equipment failure is incorporated at the subsystem level instead of global level. This study advances DES of equipment failure modeling of a truck-shovel haulage system. And also would help mine planners to comprehensively incorporate long-term maintenance plans for mobile equipment into longterm production planning.