Aim: To investigate the mediating role of eating behaviors in the relationship between chronic pain and pain catastrophizing with body mass index in people with chronic pain. Method: The research method was correlation. The sample (n = 400) were recruited through purposive sampling method from patients (males and females) with chronic pain who were visiting orthopedic, physiotherapy, neurology wards of Al-Ghadir Hospital and Farahbakhsh Clinic in Tehran in spring, summer and autumn of 2016. Participants' body mass index were collected, as well as, participants' responses on the Dutch Eating Behavior Questionnaire (VanStrien, Frijters, Bergers & Defares, 1986), Chronic Pain Questionnaire (VonKorff, Dworkin & Le Resche, 1990), and the Pain Catastrophizing Scale (Sullivan, Bishop & Pivik, 1995), then analyzed using structural equation modeling employing Amos software. Results: The results showed that the model has a good fit (df2 <3, GFI, AGFI and CFI> 0. 90 and RMSEA <0. 08). Examination of the path coefficients showed that pain (β = 0. 303, P = 0. 001; β = 0. 148, P = 0. 001) and pain catastrophizing (β =-0. 149, P = 0. 001; β = 0. 153, P = 0. 001) have a significant direct and indirect effect on body mass index, respectively. Furthermore, chronic pain, pain catastrophizing, and eating behaviors explained for 30. 5% of the variance in body mass index. Conclusion: The interaction between pain experience and pain catastrophizing with eating behaviors affects body mass index.