Crop diversification has been increasingly suggested as a sustainable approach to mitigate climate change impacts by enhancing soil organic carbon (SOC) sequestration. In intensive agricultural regions, such as the Po Valley in Italy, reliance on monocultures has significantly depleted SOC, worsening vulnerability to future climate change. To address this issue, field experiments combined with biophysical modelling (ECOSSE) and spatial interpolation techniques were used to evaluate the effectiveness of diversified cropping systems in increasing SOC under current and projected climate scenarios. This study represents the first use of ECOSSE to assess the effect of diversified systems in the Po Valley, integrating novel crop rotations with pea, reduced tillage, and circular organic amendments. It also incorporates slurry management effects and compares three complementary approaches: process-based modelling, machine learning, and spatial interpolation, under multiple general circulation models for climate scenarios. Our results indicate that crop diversification improves soil organic carbon retention substantially compared to conventional monoculture. This, in particular, benefits soils that are initially low in carbon content. However, the extent of these benefits varied considerably depending on the soil type and climate scenario. Machine learning analysis revealed temperature, rainfall, and evapotranspiration as critical features influencing simulated SOC changes. The results provide region-specific insights that can inform climate-resilient agricultural policies, including conservation agriculture incentives and payment-for-ecosystem-services schemes. These findings underline the necessity for context-specific diversification strategies to enhance agricultural resilience and sustainability in the face of climate change.<p></p>