The European Green Deal states that the European Union will have to cut 55% of carbon-dioxide (CO2) levels and implement a share of at least 40% of renewable energy sources in the electricity sector by 2030, compared to levels of 1990. While renewable energy technologies are effective in cutting CO2 levels, they also bring challenges like variability and unpredictability which affect power system stability. The TradeRES project was founded in order to create and test new innovative electricity market designs that incorporate near 100% renewable energy generation. Multiple research groups across the EU participate in TradeRES, including TU Delft and TNO. In order to be able to research future system stability, this project seeks to improve investment decision-making in the energy models EM-Lab and COMPETES, provided by TU Delft and TNO respectively. To fully utilize the sophistication of both models and to prevent development of a new model, this thesis focuses on the coupling of the models through soft-linking and its validation in a Dutch energy transition context. This thesis presents a conceptual approach and implementation towards the aforementioned soft-linking. First, the definition and requirements of soft-linking are defined to mitigate ambiguity created by previous work. These requirements are then implemented in a conceptual model describing data exchange, data mapping and timing of the models in the context of soft-linking. The conceptual model describes how the strengths of both models are utilized in the coupling. EM-Lab is an agent-based model used for studying investment behavior and has strong policy and market modules. COMPETES is an optimization model that has a very sophisticated and detailed dispatch module. The nature of these models provide coupling potential and constraints. A series of design choices describe how these models interact in the soft-linking, and how the decision was made to couple specific modules from EM-Lab with COMPETES to improve overall investment decision-making. The implementation entails the recreation of EM-Lab market modules with Python and the soft-linking of these modules with COMPETES using the software kit SpineToolbox. From EM-Lab, the CO2 market and the capacity market were recreated in Python. These modules could be used to complement COMPETES and improve overall investment decision-making. Pseudo-code of the implemented algorithms is provided and a thorough description of the programmatic interactions of the models is given. Also, the SpineToolbox implementation is shown and elaborated on. To ensure correctness, the soft-linking is validated and verified. Standard verification methods like face validity and static functional testing are used. For validation, the soft-linking is used to generate results in case studies which have been used to validate COMPETES. These case studies resemble a Dutch energy transition scenario. The results from the soft-linking are then compared to the validated COMPETES results. As the differences in results could all be explained, there is no reason to assume the soft-linking is invalid. Results of the soft-linking are provided and contain graphs regarding the CO2 market, capacity market, dispatch, and investment and decommissioning decisions. The CO2 market presents volatile and extreme results, begging the question whether the current implementation is sufficient for future power system analysis. However, the results do suggest that the soft-linking has effect on investment and decommissioning decision-making. In conclusion, additional work, like representation of mechanisms that help stabilize the CO2 prices, is required in order to use this soft-linking for future power system analysis. The potentials of soft-linking are explored and suggestions are made for future work.