Bulk air coolers (BACs) are extensively used in the deep-level mining industry to cool ambient air for underground use. BACs contribute significantly to the mine's chilled water requirements and it was therefore the purpose of this study to utilise a modelling method to predict the thermal performance of mechanically ventilated direct-contact bulk air cooling towers and perform energy optimisation. The model consisted of a combination of forward and first-principle methods based on underlying energy balance and heat transfer principles. The model allowed for the explicit estimation of the BAC outlet conditions for given inlet conditions, without the need for iterative calculations. This model was validated with operational data obtained from a bulk air cooler of a deep-level gold mine in South Africa. Average prediction errors of between 0.5 degrees C and 0.7 degrees C for the outlet water temperature were achieved. Subsequently, the model was coupled with an optimisation platform to optimise the electrical energy consumption of the BAC. Energy savings of 13% were obtained using the optimisation model for a period of high energy demand. This paper showed that such a modelling and optimisation approach could be useful to reduce the operation energy cost of thermal systems in general. (C) 2016 Elsevier Ltd. All rights reserved.