Effective weed detection is essential for efficient crop management, reduced herbicide usage, labor savings, sustainable production practices, and improved crop yields in agricultural fields. Weeds compete with plants for vital resources such as nutrients, water, space, and sunlight, which can directly impact plant growth, yield, and overall health. This research proposes a teleoperated mBot robot equipped with a camera and Ultra-Wide Band (UWB) sensors, operated using an M5Stack Core 2 module. The robot captures data from fixed anchors placed around the surveyed area. A convolutional neural network (CNN), implemented through Python code is hosted and run on an Ubuntu instance. The collected data is uploaded to Amazon S3 and EzData via a Wi-Fi network, and after analysis, the system accurately determines the location of weeds. To examine the impact of UWB sensor orientation at various range intervals, the research explored two network configurations to evaluate localization performance. Findings indicate that improved measurement accuracy is achieved by aligning the moving sensor antenna with the anchor antenna’s direction. Additionally, enhancing the number of anchors further enhances localization performance, making it more practical for optimal use. This research highlights the benefits of utilizing the proposed automated system for weed detection, which include improved efficiency, precision, scalability, and environmental sustainability compared to traditional manual weed methods. These outcomes imply that the proposed system has the potential to significantly transform weed detection practices in agricultural fields and landscapes and promote sustainable and efficient production methods.