In accelerated engineering technology and complex construction environments, unexpected events have become more frequent. To explore how engineering project employees respond innovatively, a job-crafting model was developed. This model was grounded in the theory of individual dynamic capabilities and links job crafting, improvisation, event novelty, and employee innovative behavior. To gather data, a questionnaire was distributed to 238 employees of engineering enterprises. After data collection, reliability, validity, and correlation analyses were conducted using AMOS and SPSS software. Following these analyses, linear regression analysis with the PROCESS macro tested the hypothesized paths. In addition, fuzzy-set qualitative comparative analysis (fsQCA) examined combinations of antecedent variables that led to employee innovative behavior. The regression results indicate that job crafting has a significant positive effect on innovative behavior, and that improvisation significantly mediates this relationship. Furthermore, event novelty positively moderates both the direct effect of job crafting and the mediating role of improvisation. Finally, the fsQCA identifies three configuration paths to innovative behavior, with task crafting, cognitive crafting, improvisation, and event novelty as core factors. Overall, these findings offer new perspectives for innovative problem-solving during unexpected events and provide theoretical support and practical guidance to help project managers and employees implement job crafting strategies.