---

Introduction: Cancer is the second cause of death in the world, and colon cancer is the third cause of cancer deaths. Assessment of factors which affect this cancer is important for prolonging the patient survival. The purpose of this study was to compare para-metric and Cox models as well as to deter-mine demographic and socioeconomic fact-ors affecting the 5-year survival of patients with colon cancer. Materials & Methods: In this survival stu-dy, population included 580 patients with colon cancer who were recorded in the Ca-ncer Research Center of Shahid Beheshti University of Medical Science from April 2005 to November 2006 and then were followed up for 5 years. Exponential, weib-ull, lognormal, log-logistic, and generalized gamma parametric models and Cox model were used to process the data. The models were compared with Akaike’s information criteria and Cox-Snell residual plot. Findings: Of 580 patients, 34% died. The age of patient during cancer diagnosis, gender, custody during childhood, residence location and afflicted organ were associated with the 5-year survival of patients. He calculated hazard ratios were nearly similar in different models. Based on Akaike’s inf-ormation criteria, the log-logistic and ext-ended gamma models were determined as better models, however based on Cox-Snell residual, the log-logistic and exponential models were determined as better models. Discussion & Conclusion: The result of study showed that the log-logistic model is a suitable model. Regarding the association between socioeconomic factor of residence location and 5-year survival of patients, the health promotion programs and planes is needed to improve the socioeconomic det-erminants of health status.