本文以广东省30所高校2019~2023年面板数据为样本，运用三阶段DEA测度绩效，采用fsQCA探究六类要素的组态效应。研究发现：广东高校绩效均值为0.426，纯技术效率不足是主要制约；政府支持、人才规模、科研激励是核心条件但非必要；存在四条高绩效路径（政府—人才支持型、激励主导型、人才—合作复合型、全面协同型），激励主导型覆盖度最高；科技创新能力在路径中缺席；理工类更依赖校企合作，综合类更依赖政府—人才支持。结论为分类施策提供依据。

Using panel data from 30 universities in Guangdong Province from 2019 to 2023 as a sample, this study employs a three-stage DEA to measure performance and utilizes fsQCA to explore the configurational effects of six types of factors. The findings reveal that the average performance of Guangdong universities is 0.426, with insufficient pure technical efficiency being the primary constraint. Government support, talent scale, and research incentives are core conditions but not necessary. Four high-performance pathways exist (government-talent support type, incentive led type, talent-cooperation composite type, and comprehensive coordination type), with the incentive-led type having the highest coverage. Scientific and technological innovation capability is absent in these pathways. Science and engineering disciplines rely more on university-enterprise collaboration, while comprehensive disciplines depend more on government-talent support. The conclusions provide a basis for differentiated policy implementation.