Background and Aims: Soil net nitrogen mineralization (Nmin) rate and its temperature sensitivity (Q10) are critical parameters for predicting nitrogen availability under global warming. However, the controlling factors of the net Nmin rate and its Q10 at the regional scale remain highly uncertain, especially in different ecosystem types. Methods: An incubation experiment at three different temperatures was conducted to measure the net Nmin rate and to calculate its Q10. High-throughput sequencing was applied to explore the microbial community diversity and composition. Results: Results showed that the overall net Nmin rate was highest in paddy fields, followed by forest and dryland fields, whereas its Q10 was not different among ecosystems. Structural equation models revealed that the net Nmin rate was mainly affected by soil properties in dryland habitats, whereas it was mainly influenced by soil microbial communities and mean annual temperature (MAT) in paddy habitats. In forest habitats, the net Nmin rate was mainly affected by bacterial communities rather than fungal communities. The Q10 of net Nmin was close to 2.0, indicating that the net Nmin rate is sensitive to climatic warming. The Q10 was impacted by the bacterial community composition in dryland soils, whereas it was influenced by soil properties and bacterial diversity in paddy and forest soils. Conclusion: This study provides a novel quantification on the contribution of soil microbial diversity and community composition to soil net Nmin process in three ecosystem types, with implications for microbial-mediated N mineralization process under climatic change scenarios. [ABSTRACT FROM AUTHOR]