Predictive maintenance employing machine learning techniques and big data analytics is a benefit to the industrial business in the Industry 4.0 era. Companies, on the other hand, have difficulties as they move from reactive to predictive manufacturing processes. The purpose of this paper is to demonstrate how data analytics and machine learning approaches may be utilized to predict production delays in a quarry firm as a case study. The dataset contains production records for six months, with a total of 20 columns for each production record for two machines. Cross Industry Standard Process for Data Mining approach is followed to build the machine learning models. Four predictive models were created using machine learning algorithms such as Decision Tree, Neural Network, Random Forest, and Nave Bayes. The results show that Multilayer Perceptron Neural Network outperforms other techniques and accurately predicts production delays with a sensitivity score of 0.979. The quarry company's improved decision-making reducing potential production line delays demonstrates the value of this study.