The thesis focuses on the problem of Impact Monitoring of Near-Earth Objects (NEOs) and proposes a novel approach using artificial neural networks. The primary objective is to develop a neural network capable of determining the potential danger posed by an observed NEO by predicting its minimum orbital intersection distance (MOID) based on its coordinates at a specific epoch. While traditional methods for computing MOID and impact probability exist, the thesis aims to explore the feasibility of leveraging machine learning techniques to expedite the preliminary assessment of asteroids, potentially saving time and resources in the process. By harnessing the power of neural networks, the research seeks to enhance our ability to identify and monitor potentially hazardous NEOs, contributing to the field of Impact Monitoring and risk assessment in near-Earth space.