The urban heat island (UHI) effect describes elevated temperatures in urban areas relative to surrounding rural regions. Reductions in green spaces and natural cooling sources exacerbate this phenomenon, with prior studies demonstrating that urban modifications substantially contribute to UHI intensity. Therefore, designing a digital tool to identify and understand how new modifications affect temperature conditions plays a crucial role in optimized city planning. The digital twin (DT) concept can be considered as one approach to address this scenario. A DT is a digital platform that works as a one-to-one mapping between a digital simulation and a real-world environment. Our design science research project proposes a novel solution for this scenario using a DT system based on the city of Lahti, Finland. Our solution, a DT system, allows users to implement modifications, including trees, built-ups, and water bodies in the Lahti city area, to forecast their influence on land surface temperature (LST). The DT is combined with a machine learning (ML) model, which was trained to predict the LST, focusing on factors including geographical locations and spectral indices. The LST is forecasted, considering the changes that occurred after the modifications. Our solution is evaluated through a technical evaluation of the ML algorithm and a user evaluation with environmental specialists, urban planners, and students of sustainable urban planning studies.