Massive and expensive mining machines, which are used for overburden stripping, play a major role in the coal mine. One of the prominent mining machines is dragline which has a vital role in the production phase. In this sense, any stoppage in its operation leads to an extreme production rate reduction and consequently higher operating cost. The conventional methods that are widely applied for mining machinery maintenance programs are run-to-failure scenario and preventive maintenance action since they are the simplest maintenance programs. However, the predictive maintenance scenario is the interest of machine owners and operators. Regards to machine owners' satisfaction, any methodology that provides predictive maintenance should be considered. In this paper, one of the key predictive maintenance methods, which is machine learning as a cutting-edge of the predictive maintenance, was considered. The objective of this paper is to classify the failure modes of the walking dragline regards to develop a maintenance strategic plan. In order to carry out the training process, some major pre-processing actions have been applied. The utilized methodologies are Multi-layer Perceptron, Radial Basis Function, K-Nearest Neighbours and Enhanced K-nearest Neighbours algorithms. The applied algorithms predict the most failure prone mode. Based on the assessment of the used algorithm, enhanced K-Nearest Neighbours revealed better result compare to others, which shows 82 percent of training accuracy. The novelty of the study is based on the application of machine learning methodologies, which lead to reliable mining operation in the context of the mine automation.