As smart grids develop, power systems become more complex, and the role of data gain considerable importance for the reliability of power supply. Thus, data processing techniques have to be investigated and compared to increase the efficiency of asset management decisions. In this paper, we explore several black-box models in order to predict power exchanges through a high-voltage direct-current line between Sweden and Denmark, using publicly available data on loads and power prices. An auto-regressive moving average with external input model based on load data provides the most accurate forecasts according to mean square error and other selected criteria. This is the first step to build a more comprehensive model that will also include other technical data such as maintenance and unplanned outages, but also macroeconomic factors. The final goal is to provide network operators with a parsimonious sequential model composed of several modules giving accurate predictions that support efficient asset management decisions.