The present research paper deals with AlSi10Mg alloy/MWCNT metal matrix composite brake pads with varying weight percentages. The Development of new brake pad materials has been done at compacting load (10 Tons) and sintering temperature (450 °C) using the powder metallurgy process. The input parameters of 40 N normal load, 500 rpm, 150 °C pin temperature as the near-optimal combination of parameters for minimum wear and maximum coefficient of friction compared to other test conditions have been observed by the design of experiment (DOE). The overall average percentage error in the output against experiment output is less than 1%. Analysis of variance (ANOVA) indicates that load is a most significant factor than speed and temperature for wear and CoF. Artificial Neural Network (ANN) is used to develop a prediction model to calculate wear and coefficient of friction for different loads, pin heating temperature and speed. The model developed shows a strong correlation with experimental output. The experimental and predictive model developed from artificial neural network are strongly correlated with a correlation factor of 0.99447 for the training algorithm of Levenberg-Marquardt technique. ANN prediction model and Taguchi L16 array reveal that the experimental and predicted data for mass loss and CoF have less than 3% and 4% error, respectively. The closeness between the artificial neural network and experiment results enhances the scope of ANN for predicting the wear of materials. The model will help engineers to predict the failure of components with reference to running time and can be applied in automobile or manufacturing sectors to study wear, thus saving their time and cost in carrying out experiments.