Providing accurate and up-to-date agricultural vegetation maps is a very important task for agricultural land evaluation and monitoring. These maps allow various kinds of spatial analyses could be conducted to optimally manage and utilize of land resources. One of the newly developed approaches in information extraction from remote sensing data is objectbased approach or widely known as Geographic Object-Based Image Analysis (GEOBIA). This study aims to utilize GEOBIA and a pan-sharpened WorldView-2 image (0.5 m pixel size) to identify and map agricultural vegetation types in part of Dieng Plateau, Central Java, Indonesia. A multiresolution segmentation algorithm was used to partition the image into vegetation object candidates based on some segmentation criteria. The accuracy of segments created were evaluated by visually comparing the segmentation result with the objects border on the image and field visit. A hierarchical conceptual model was created to systematically classify targeted agricultural vegetation objects, and the relevant interpretation keys for each object were identified. For the classification process we implemented a rule-based classification based on segment’s values, shape, homogeneity, texture, compactness, asymmetry, roundness, elliptic fit, number of pixel and border length. The result showed that the combination of GEOBIA and WorldView-2 were able to discriminate and map the types of agricultural vegetation into cabbage, carica, carrot, chili, potato, potato with soil solarization, and tamarillo with a reasonably high accuracy.