Dengue Hemorrhagic Fever (DHF), an acute capillary leak syndrome caused by dengue virus, is one of the major public health problems in Indonesia. Several major DHF outbreaks have occurred, and the incidence rate has increased from year to year. The increasing use of social media for health-related activities and the geolocation social media data can be leveraged to enhance the current surveillance system for controlling communicable diseases, including DHF. This study explores Instagram, one of the most popular social media in Indonesia, to analyze posts about DHF toward improving the surveillance system. We extract Instagram posts m 2017-2018 with geolocation and date-tune value, that contained one or more DHF keywords in Indonesian language (i.e., demam berdarah dengue, demam berdarah, dengue. Aides aegypti, and fogging), resulting in 665 posts in dataset 2017 and 976 m dataset 2018. The preprocessmg of the dataset includes tokemzation, word normalization, stop word removal, stemming, and building the term-document matrix. The training dataset is labeled mto news, education and information, and other classes, and the K-nearest neighbors (KNN) is applied to classify the DHF posts. The accuracy of the KNN classifier is 71% for the 2017 dataset and 77% for the 2018 dataset. For the deployment, we perform spatial analysis and compare the results with the actual DHF cases from the Ministry of Health of Indonesia. The spatial analysis results report that 17 provinces reveal similar changes between the number of DHF posts on Instagram and the number of actual cases from 2017 to 2018, while 17 other provinces reveal the opposing changes. The results and findings of this study show that Instagram posts may contribute to enhance the current surveillance system. For example, the frequency of Instagram posts and its sharp increase may indicate unusual events about the DHF situation in a specific province.