Nowadays, vegetation classification can be used to find out the latest information about the characteristics and distribution of vegetation in an area. However, a conservative process to differentiate vegetation was ineffective. Some of those limitations are poor accessibility that does work less safety, time-consuming, and needs a lot of human resources. On the other hand, remote sensing offers solutions that cannot be done by the simple method, such as how to take the data, time-consuming are less, and human resource needs are less as well. The purpose of this study was to classify, measured the area of each vegetation, and compared the effectiveness of the unsupervised used K-Means algorithm and supervised used Object Base Image Analysis algorithm methods vegetation classification. For accuracy calculation with confusion matrix, the classification results of the two methods were compared with the manual digitization method. Data was taken using drones in the area of the Curah Manis Sempolan Nature Reserve 1. Classification of vegetation consists of 5 vegetation types, which was apak, bush, pine, bendo, and dadap. The total area of the study area was 1.633 ha, and area vegetation of each classification was apak 0.224 ha; bush 0.748 ha; pine 0.394 ha; bendo 0.222 ha; and dadap 0.045 ha. The results of the calculation of accuracy showed that the unsupervised method had a value for overall accuracy of 80% and kappa accuracy of 73.58%. Then, in the supervised for overall accuracy is 68% and kappa accuracy of 58.72%. Keywords: classification, drone, remote sensing, satellite