The agriculture industry plays a crucial role in a country's economic growth cycle across the world. In the field of agriculture, weeds are the major concern for farmers as they reduce crop yield. They compete with the crops vigorously for water, sunlight, and soil. Thus, it is important to remove these weeds at early growth stages. As a result, weed classification is required, but it is a time-consuming and difficult effort to do it manually. An automated weed system based on images is required for weed classification. Deep learning approaches show better performance in image classification, among which Convolutional Neural Networks are the most preferable. Therefore, in this study, the classification of nine different species of weeds is performed using six pre-trained convolutional neural networks: Xception, Inceptionv3, VGG16, VGG19, AlexNet, and InceptionResNetv2. These pre-trained models' performance parameters, such as accuracy, f1-score, precision, and recall, are computed. The results suggest that Xception outperforms the other networks evaluated, with an accuracy of 94%.