Abstract: With the intensification of global climate change, the urban heat island effect has become an increasingly prominent environmental issue in cities worldwide. Urban green spaces are widely recognized as an effective ecological strategy for mitigating urban thermal environments. Although numerous studies have examined the cooling effects of urban green spaces from multiple spatial scales, limitations remain in the consistency of scale classification, the integration of analytical approaches, and the comprehensive understanding of cross-scale mechanisms, which hinder the effective application of research findings in urban planning practice. This study provides a systematic review of recent literature on the cooling effects of urban green spaces, focusing on research scales, methodological approaches, influencing factors, as well as current limitations and future directions. The results indicate that existing studies can be broadly categorized into three spatial scales: street scale, patch scale, and landscape scale. At different spatial scales, the characteristics of green spaces, cooling processes, and dominant mechanisms exhibit significant differences, reflecting clear scale-dependent patterns. In terms of methodology, four major approaches are commonly employed: remote sensing, field observations, mobile measurements, and numerical simulations. Remote sensing is widely used to analyze spatial patterns of land surface temperature and assess cooling effects over large areas. Field observations and mobile measurements provide detailed microclimatic data, enabling the investigation of fine-scale thermal processes. Numerical simulation models further support the exploration of thermal mechanisms and the evaluation of planning scenarios. The analysis of influencing factors shows that the cooling effects of urban green spaces are jointly determined by intrinsic characteristics of green spaces, spatial configuration, and surrounding urban environmental conditions. Moreover, the relative importance of these factors varies significantly across spatial scales, indicating strong scale-dependent interactions. Despite substantial progress, several limitations remain in current research. Many studies rely heavily on medium-resolution remote sensing data, which may introduce mixed-pixel effects in complex urban environments. In addition, studies conducted at a single spatial scale often lack cross-scale integration. Future research should therefore emphasize the integration of high-resolution and multi-source data, the development of cross-scale analytical frameworks, and the incorporation of human thermal exposure and thermal comfort into urban planning applications. These efforts will provide stronger scientific support for climate-responsive urban green space planning and design.