Quality machine learning that has benefits for precision agriculture, agriculture that is the main goal is how the technique of providing nutrients, fertilizers and water follows the needs of plants through sensors as an indicator to find out the needs of the plant, in this case oil palm nurseries. Problems that often arise in the process of oil palm seedling are the waste of water resources, fertilizers, nutrients and inaccurate application of water PH. Excessive application of chemical fertilizers and leftovers not consumed by plants can damage the soil structure to become barren and nutrient-poor. The research is basis on embedded systems while still using 3–5 variables to describe the needs of the plant, through the affinity data approach as the basis for making data labeling for the initial preparation of machine learning. The labeling in question can be using to predict the need for oil palm seedlings with watering time action based on affinity data. Through Data Affinity, data to determine the similarity of data between the expected value and the value obtained from the sensor. In machine learning, 10 variables can be calculating to consider the needs of oil palm seeds.