Landscape environment plays a significant role in the innovation development of resourcebased cities. From the perspective of eco-city, this study selects 23 regenerated resource-based cities in China as the research area and builds an interpretable machine learning method to uncover how service facilities, social economy, and the landscape environment relate to urban innovation output. Firstly, a reasonable factor framework was constructed. Secondly, the influencing factors were screened and models were compared. Thirdly, the contribution degree of factors was analysed using SHAP. Fourthly, the interaction effect of factors was analysed using PDP. This study reveals that science and education venues exert a significant positive effect on the innovation output of regenerated resource-based cities. High-quality landscape environments can effectively stimulate the innovation vitality of these venues. Explainable machine learning method can incorporate the landscape environment into the urban research framework, revealing the relationships between the landscape environment and various urban features. Through the interpretation of the results, targeted development suggestions can be provided for the city, which has strong practical guiding significance.