Labels are key to an accurate image prediction. This study combines image classification with decision tree classification to reduce soil identification period, solidify soil analysis result identification and help farmers identify the appropriate crop to grow on specific soil pH levels. To determine the recommended pH levels for the major crops grown in Abra, Philippines, and to determine the logical sequence of label-image.py, the study used the University of West Florida guide to conduct and write a literature review. To develop a decision-tree based classification system for crops, the study used the principle of splitting criteria. The idea is to partition the predictor space to such a degree that the reply variable’s standards are alike. Out of the 19 crops, 32% grows within 5.5-6.5 range like Taro, Palay, Eggplant, Melon, Rambutan, and Yams, while 5% each thrives in the 4.7-5.7 range, 6.5-7.5 range, and 6.5-8.0 range like Sugarcane, Gourd, and Beans respectively. Crops have different soil pH level requirements to grow. Predictions of label_image.py are sorted. Thus the predictions are easily understood. Tree-based classification can be easily set up with an image classifier. It helps in determining the appropriate crop to be planted on a specific pH level.