Phytophthora infestans, causing late blight, is a widespread and destructive disease on tomatoes, resulting in severe yield losses and economic impacts. This study proposes an integrated IoT system for detecting late blight and continuously monitoring tomato plants. The project utilized modern image processing, using the Convolution Neural Network (CNN), Artificial Neural Network (ANN), and ResNet50 Models to identify captured images from tomato plants. There is also an environmental monitoring system based on sensors of critical disease-development variables, such as humidity, temperature, and soil moisture, complementary to the image analysis. The proposed system further helps cast timely alerts to the farmers, enabling timely intervention to reduce the spread of disease. By integrating Machine Learning models with environmental information, this method seeks to elevate agricultural productivity, decrease crop losses, and facilitate sustainable agricultural techniques by enabling proactive disease management. Experimental results demonstrate that ResNet50 outperformed other models, achieving an accuracy of 95%, precision of 100%, recall of 90%, and F1-score of 92.5%, while (CNN) achieved accuracy of 93%, precision of 100%, recall of 86%, and F1-score of 92.5%, the (ANN) result demonstrated accuracy of 68.3%, precision of 91%, recall of 40.7%, and F1-score of 56.2%. These results highlight the effectiveness of integrating deep learning with IoT in precision agriculture for early disease detection and proactive crop management.