Journal of Intelligent Computing Qian Ding 16 4 2025 https://doi.org/10.6025/jic/2025/16/4/137-144 https://www.dline.info/jic/fulltext/v16n4/jicv16n4_1.pdf This work explores the relationship between employee performance commitments and corporate earnings management, particularly in the context of mergers and acquisitions (M&A). It highlights that while performance commitments can align employee goals with corporate strategy, reduce information asymmetry, and signal market confidence, they may also incentivize earnings management when targets are unrealistically high. The study leverages a decision tree model based on information entropy to analyze this dynamic using M&A data from 2009 to 2022. The model uses real earnings management as the dependent variable and includes commitment duration, commitment amount, and several financial controls. Results indicate a positive correlation between unrealistic performance commitments and heightened earnings management. A case study of "XX Smart Enterprise" illustrates how excessive commitments led to significant financial strain and potential earnings manipulation. The paper proposes an optimized commitment level of approximately 30% to mitigate such risks. By integrating a generalized entropy increase algorithm into the traditional decision tree framework, the model improves prediction accuracy and offers practical guidance for setting realistic performance targets. The findings caution against overly ambitious commitments in M&A agreements and emphasize the need for balanced incentive structures to prevent short term earnings manipulation that could harm long term corporate health and stakeholder trust.