Stock price prediction has long been a complex and challenging task due to its non-linear patterns and high volatility. Despite of its uncertainty, there is a continuous increase of investors in the share market. So, it is the need of the hour for a deep investigation of the market behavior and to develop robust model with better accuracy along with explainability. Although many machine learning approaches have been made to unfold the behavior, there lies a trade-off between accuracy and interpretability or explainability of the models. This paper presents a deep learning approach using a neural network to forecast the closing stock price of Ultratech Cement Limited based on historical trading data. Here is an implementation and evaluation of a multilayer perceptron (MLP) architecture using normalized market indicators. Further, the model’s interpretability is addressed using SHAP (SHapley Additive exPlanations) to uncover the influence of each individual feature. The approach achieves high accuracy on test and unseen datasets, and SHAP insights reveal the impact of features, enhancing transparency of the model. This study offers a scalable and explainable framework for financial forecasting.