To understand forest dynamics under today’s changing environmental conditions, it is important to analyze the state of forests at large scales. Forest inventories are not available for all regions, so it is important to use other additional methods, e.g., remote sensing observations. Increasingly, remotely sensed data based on optical instruments and airborne LIDAR are becoming widely available for forests. There is great potential in analyzing these measurements and gaining an understanding of forest states. In this work, we combine the new-generation radiative transfer model mScope with the individual-based forest model FORMIND to generate reflectance spectra for forests. Combining the two models allows us to account for species diversity at different height layers in the forest. We compare the generated reflectances for forest stands in Finland, in the region of North Karelia, with Sentinel-2 measurements. We investigate which level of forest representation gives the best results and explore the influence of different calculation methods of mean leaf parameters. For the majority of the forest stands, we generated good reflectances with all levels of forest representation compared to the measured reflectance. Good correlations were also found for the vegetation indices (especially NDVI with R2=0.62). This work provides a forward modeling approach for relating forest reflectance to forest characteristics. With this tool, it is possible to analyze a large set of forest stands with corresponding reflectances. This opens up the possibility to understand how reflectance is related to succession and different forest conditions.