For an energy management system (EMS) of a charging station (CS), information on future load is crucial. Existing models primarily focus on load forecasting for large charging stations. In this study, three different load forecasting models based on real data from a public CS with two charging points are developed. The models include two persistent models and one model that utilizes a machine learning algorithm. To assess the impact of forecasting accuracy on operational costs, a case study with dynamic electricity prices and a stationary battery storage is conducted. Using the load predictions, a mixed-integer linear programming problem is formulated to optimize the scheduling of the stationary battery charging.