The construction industry is recognized as one of the most hazardous sectors, necessitating a deeper understanding of cognitive and individual factors influencing workers' safety performance. This study investigated the relationship between the number and level of job skills with hazard identification and safety risk perception among construction workers. The research was conducted on 162 workers employed in construction projects in Shiraz, Mashhad, and Tehran during fall 2024. Demographic data, number of specialties, and skill levels were collected using a researcher-developed questionnaire (Cronbach's alpha = 0.88). Hazard identification was assessed using real workplace images, and safety risk perception was evaluated using Tixier et al.'s standard instrument. Findings revealed that 74.7% of workers were single-skilled and 87% had received no safety training. While age, marital status, and nationality showed no significant relationship with skill level, education demonstrated a direct significant relationship (p = 0.003). Safety training significantly impacted both hazard identification and risk perception (p = 0.000). Age showed no significant relationship with safety variables, whereas education demonstrated significant relationships with both. Positive correlations were observed between number of skills and hazard identification (r = 0.549, p = 0.000) and risk perception (r = 0.444, p = 0.000). Similarly, significant correlations were found between skill level and hazard identification (r = 0.387, p = 0.000) and risk perception (r = 0.258, p = 0.001). The Sobel test confirmed the mediating role of hazard identification in the relationship between number of specialties and risk perception (z = 2.811). It is concluded that multi-skilled workers with higher skill levels possess greater ability to identify hazards and perceive safety risks, with hazard identification serving as a mechanism linking increased skill variety to improved safety risk perception.