The conventional methods for determining the shear strength parameters of soil, namely cohesion (C) and angle of internal friction (φ), involve time-consuming and expensive machinery. Also, the extraction metal ore processing into metal and machine manufacturing involves a high level of carbon emission. During the operation of these machines a large quantity of electricity is generated in thermal power plants, leading to an indirect increase in the carbon footprint, thus suggesting a need for the adoption of more sustainable practices. This study is aimed at reducing net zero emission by laboratory testing machines by develop a predictive machine-learning model for estimating soil shear strength parameters by utilizing the existing soil data. An artificial neural network (ANN) is one such model inspired by the human brain, with the ability to learn complex patterns and relationships in data. For this, 88 soil samples in all were gathered from the literature that was available. Basic index properties of soil are used as inputs and C, φ are predicted as outputs. A neural network is developed using a Bayesian regularization optimization algorithm. The model developed is evaluated using the root mean squared error, coefficient of determination, and mean absolute error. The efficacy demonstrated by the ANN model ensures reliable predictions, thus promoting the adoption of these predictive models, leading to net zero carbon emissions in the testing field of geotechnical engineering.