In modern higher education institutions, the effective evaluation of research performance and the development of incentive-based compensation mechanisms based on such evaluations have become strategic priorities in human resource management. Traditional performance evaluation models primarily rely on quantitative indicators and therefore fail to adequately capture essential dimensions of research performance, such as quality, societal impact, and disciplinary complexity, while also increasing the risk of subjective decisionmaking. In this context, the application of artificial intelligence (AI) technologies and multi-criteria decisionmaking (MCDM) methods offers new opportunities for improving research performance management systems. The purpose of this study is to develop a conceptual model of an AI-supported differential compensation system that ensures the objective, transparent, and adaptive evaluation of research performance in higher education institutions. The proposed methodological framework is based on multi-criteria decision-making theory and integrates the Analytic Hierarchy Process (AHP), Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), the Entropy Weight Method, and weighted linear aggregation models. Within this framework, research performance indicators are normalized, their relative weights are determined through a combination of expert judgments and AI algorithms, and a comprehensive multi-criteria performance index is constructed. The findings indicate that the proposed model enables a comprehensive assessment of research performance by incorporating not only quantitative outputs but also qualitative and societal impact indicators. The integration of AI algorithms allows real-time updating of performance indicators, considers disciplinary differences, and facilitates the adaptation of compensation mechanisms to institutional strategic priorities. This approach enhances academic motivation, strengthens objectivity in decision-making, and contributes to the effective management of the research potential of higher education institutions. In conclusion, the proposed AI-based multi-criteria performance index provides a methodologically sound and practically applicable innovative framework for evaluating research performance and designing differential compensation systems in higher education institutions.