The increasing challenges in modern agriculture—such as population growth, climate change, and limited natural resources—have prompted the development of intelligent solutions for enhancing crop production efficiency. Conventional greenhouse microclimate control systems typically rely on environmental sensors and predefined logic rules, often failing to adapt dynamically to the biological needs of plants throughout their growth cycle. This paper introduces an intelligent microclimate regulation system that leverages morphological analysis of plants through machine learning and an Internet of Things infrastructure. The proposed system utilizes a camera connected to a Raspberry Pi to periodically capture images of plants. These images are analyzed using convolutional neural networks to classify the current growth stage of each plant. Based on the classification result, a microcontroller adjusts the operation of actuators—such as heating, ventilation, humidification, lighting, and carbon dioxide enrichment—to create optimal growing conditions. A hybrid dataset was used for training and evaluation, consisting of open-source and experimentally collected images under varying lighting conditions. Three classification methods were implemented and compared, a custom CNN model, MobileNetV2 with transfer learning, and a support vector machine using a histogram of oriented gradients descriptor. The CNN achieved the highest accuracy at 88.2%, outperforming the other models. The study demonstrates that morphology-based, vision-driven control offers a promising alternative to conventional sensor-only methods, enabling context-aware and biologically informed climate adjustments.