When classifying satellite imagery cover is usually preceded by a supervised or unsupervised classification, directly from the system software program automatically dividing several clusters. However, especially for land cover that has similarities, it should be further classified by supervised classification or with human assistance by making a sample of land cover first. If not, there will be an error which is often referred to as omission or commission, namely omission, which means that class x omission belongs to another class, while the commission for adding class x is included in one of the land cover classes. The study tried to observe the spectral changes of both land cover every week for 7 weeks from rice plants at the age of 4 weeks to 11 weeks before harvest. By observing the reflectance every Saturday in the morning with a reflectometer, it can be seen that the two land coverages between rice fields and football field grass are similar. When lowland rice plants are still young, their reflectance position is higher than field grass, but then the reflectance position of field grass is always higher. The results of the observations were initially the red-green-blue position for rice fields 103-124-112 and for field grass 79-100-69. Furthermore, at the end of the observation before harvest, when the rice was 11 weeks old, the position of the rice fields was below it, namely red-green-blue for fields 44-63-32 and for field grass 78-94-43. In this case, it is recommended that for both land cover, a separate sample set is made between lowland rice and field grass, namely by marking the visual characteristics in the field, namely by distinguishing elements such as hue, texture, size, pattern, shadow, site, and very important association.