Dengue fever is a one of the dangerous diseases and very often causes casualties every year, especially in the tropics or subtropics countries. Dengue fever cases increase during the rainy season, many factors affect the spread of dengue fever, such as vegetation, population and landfills. The aim of this research is to predict the number of cases of Dengue fever using support vector machine. The data used are dengue data in Bandar Lampung City, weather data, population data and distance matrix data between dengue fever events with each other. The amount of data used is 1,080 data with 3 kernels: linear, gaussian and polynomial. In this study four experiments were carried out, the first two experiments were carried out without Feature Selection and the next two experiments were carried out with Feature Selection. After the experiment was found The best performance in the experiment with Feature Selection with 44 Variables. From the experiments conducted, Gaussian kernel achieved the highest R 2 value which is 75.52%, while the Linear kernel and Polynomial Kernel achieved R 2 value of 74.61% and 75.15%, respectively.