Modern lighting solutions have become smart due to saving more energy and advancing technology like introducing light-emitting diodes (LEDs). LEDs offer many new possibilities on how to interact with lighting in order to provide the opportunity for smart lighting systems. Smart lighting systems combined with sensors automate the lighting process and can go beyond illumination, affect productivity, well-being, and emotions. Multiple applications exist in smart cities, warehouses, and even residentials. This Bachelor Thesis aims to examine current methods and to create a predictive control system for indoor lighting using smart lights from Philips Hue in combination with a Raspberry Pi and multiple sensors (motion, temperature, and lighting). We will also take a look at the technical and methodical background. For the experiment, we will collect user data and use it with a machine learning algorithm (Very Fast Decision Tree) to predict lighting in a real-life environment. At the end of the research, we are going to evaluate the system’s accuracy by measuring the decisions with time metrics. After a sensitivity analysis, we can say that the evaluated system can make some inaccurate predictions due to changed environmental values like longer daytime and different weather status (e.g., less cloudy weather).