Purpose: In this study we tried to evaluate the predictive factors for survival in patients with upper urinary tract tumors. Materials and methods: From 1993 to 2003, 46 patients were treated by standard nephroureterectomy for upper urinary tract tumor, but only 24 patients (52%) who had regular follow-up were included in the study. Age, sex, presenting symptoms of the patients, tumor localization, tumor stage and grade were analyzed with respect to survival. Univariate and multivariate analyses were done using Kaplan-Meier method with log-rank test and Cox proportional hazards regression model, respectively. Results: The median of patient age was 61 years (34-74). Of the 24 patients, 9 (37.5%) were disease-free and alive at a mean time of 54 (26-97) months, 8 (33.3%) died of disease at a mean period of 23.4 months (2 because of bladder tumor, 2 had liver metastases, I had lung metastasis and 3 had lung and liver metastases) and 7 (29.2%) died disease-free at a mean period of 30.3 months. Metastases were detected in a mean period of 11.8 (6-24) months. Survival according to tumor stage Ta, T1-2, and invasive tumors were 87.5, 43.9, 15.7 months (p=0.0001), respectively. Survival of the patients with low-grade tumors was significantly longer than those with high-grade tumors (77.3 and 31.4 months, respectively, p = 0.01). Patients with pelvis tumors when compared to ureter tumors (28.5 and 61.6 months, respectively, p = 0.038) and those presenting with flank pain when compared to those presenting with macroscopic hematuria and bladder cancer (17.7, 45.7, and 57.9 months, respectively, p=0.046) had shorter survival rates. When multivariate analyses were done using Cox regression test, the only factor that affected survival was the stage of the tumor. Age and gender had no impact on survival. Conclusions: In univariate analysis, the stage, grade, localization of the tumor and presenting symptoms were found important predictors that affect the prognosis of the transitional carcinoma of the upper tract. However, tumor stage was the only independent predictor of survival in multivariate analysis. For high grade and high stage tumors, really effective adjuvant treatments along with aggressive surgery may be considered.