Against the backdrop of China’s higher education shifting from elite to mass education, coupled with post-COVID-19 economic downward pressure, college students face unprecedented employment challenges, making entrepreneurship a crucial solution. However, their entrepreneurial success rate remains below 10% due to insufficient experience and environmental uncertainties, highlighting the urgency of scientific risk assessment for undergraduate entrepreneurial projects. Existing research has explored neural network-based assessment methods, but these suffer from slow convergence and overfitting issues, failing to meet practical evaluation needs. To address this gap, this study first constructs a risk assessment index system for undergraduate entrepreneurial projects—covering 6 primary dimensions (college students and their team risk, entrepreneurial project risk, entrepreneurial technology risk, entrepreneurial industry market risk, entrepreneurial policy and industrial risk, entrepreneurial project regional influence) and 29 secondary indicators—by integrating prior research principles and current economic downturn characteristics. Second, it proposes a novel assessment model based on the cloud model, which effectively handles the ambiguity and randomness of risk factors. Finally, the feasibility and effectiveness of the proposed model are verified through empirical tests on four actual undergraduate entrepreneurial projects. This research not only enriches the theoretical framework of undergraduate entrepreneurial risk assessment but also provides practical decision support for venture capital institutions and entrepreneurial management departments.