In verticultural planting, which is a method of growing crops in narrow land, especially in the seedling stage until the plants can be harvested, watering the plants is something that must be considered because it requires special treatment, which must be careful, so that plant seeds are not lost or come out of the planting container at the seeding stage, and the plants are not damaged until
the plants can be harvested due to uncontrolled watering and based on scheduling only, so that plant growth is less than optimal. Therefore, in this research, an automatic watering system was designed to adapt to the environmental conditions around the plants. Automatic control of watering is
determined using the backpropagation neural network algorithm that can decide whether to water or not by activating or deactivating the relay based on the input
values of temperature, air humidity, and soil moisture obtained from the DS18B20, DHT11, and soil moisture (YL-69) sensors processed on the Raspberry Pi 4. Data from sensor readings and watering status can be monitored via a web-based application. The calculation of the backpropagation algorithm produces an accuracy rate of success in determining watering at 93.33% with an error percentage of 6.67%.