The cold temperate forest ecosystem is a crucial ecological zone in China, significantly impacted by human activities. To understand the impact of restoration on soil microbial communities following disturbance, this study employed high-throughput sequencing technology to systematically examine the assembly patterns and processes of soil microbial communities under two restoration modes (nature restoration (NR) and artificial restoration (AR)) in this forest ecosystem. The results indicated that the concentrations of total nitrogen (TN), alkaline hydrolysable nitrogen (AN), dissolved organic carbon (DOC) and soil organic carbon (SOC) were significantly higher in soils under natural restoration compared to artificial restoration. The α-diversity of soil bacteria remained unchanged, while soil fungal α-diversity changed significantly across different restoration modes. Furthermore, different restoration modes significantly alter the β-diversity of soil microbial (bacterial and fungal) communities. The relative abundance of soil microbial (bacterial and fungal) changed significantly across different forest restoration strategies, i.e., the relative abundance of increased in natural restoration compared to that in natural forest; similarly, both and showed higher relative abundances in natural restoration than those in natural forest. Furthermore, analysis of variance for differences between groups shows that plays a crucial role in artificial restoration. Community assembly analyses indicated that that soil microbial (bacterial and fungal) communities were primarily driven by deterministic processes in both restoration models. In short, our study improves our comprehension of how soil microbial communities respond to different restoration methods in temperate forest ecosystems, providing valuable insights for their sustainable management.