Accurate plant classification by leaves is vital in precision agriculture for monitoring and managing crop health. This study evaluates the performance of VGG-16 and VGG-19 convolutional neural networks in classifying agricultural plants by their leaves. We evaluate both models for four crop types: Cashew, Cassava, Maize, and Tomato on a larger dataset. The high classification ability of the VGG-16 model can be illustrated in relation to Cassava (Precision: 96.3%, Recall: 92.8%) and Tomato (Precision: 98.4%, Recall: 87.9%), while the accuracy is low for Maize (75.2%). On the other hand, when we look at VGG-19 model results, it performs marginally well in the case of Cashew (96.4%) and Tomato (98.7%), but comparatively better for Maize with high precision (97.4%) and recall (94.4%). The VGG-19 deep learning model thereby provides improved precision in classification as well as robustness for precision agricultural applications, according to our results. The proposed study demonstrates how advanced deep learning techniques can be utilized to enhance agricultural practices through effective monitoring and management of crops.