Weeds present a substantial challenge to agriculture, causing significant yield losses and elevating management costs. This study explores the automation of weed detection in crops through the deployment of advanced deep learning models. Two prominent models, YOLOv8 and InceptionV3, were extensively evaluated for their performance in accurately identifying weeds within agricultural imagery. The YOLOv8 model has proved to be the most effective method by achieving absolute accuracy in the differentiation of crops from weeds. The findings of the study highlighted the efficiency of weed management systems based on the YOLOv8 and its potential towards the sustainable development of agriculture through better weed control systems and then overall crop protection and resource optimization.