Objective: To investigate the relationship between the marginal width of transurethral resection of bladder tumor (TURBT) and bladder cancer recurrence to define the ideal resection marginal width. Methods: Medical records of 143 patients with non-muscle invasive bladder cancer were retrospectively reviewed. The patients were divided into three groups according to the width of resection margin. All patients were followed for a minimum of 24 months to investigate the timing of tumor recurrence and recurrence in situ. Log-rank test, Cox regression, and Kaplan-Meier estimator were performed to compare the three groups, respectively. Results: Tumor size, primary/recurrent tumor, pathological grade, resection marginal width, and margin status affected the tumor recurrence. The same factors also affected the tumor recurrence in situ. The recurrence and the recurrence in situ rate of 10 mm marginal width group were significantly higher than 15 mm and 20 mm marginal width groups (P = 0.005); similarly, the recurrence and the recurrence in situ rate in patients with positive margin were higher than those with negative margin (P < 0.001). The postoperative complication rate in 20 mm marginal width group was significantly higher compared with the 10 mm and 15 mm groups (P = 0.041). Conclusions: The positive margin of TURBT is associated with increased risk of both tumor recurrence and recurrence in situ. We recommend a marginal width of 15 mm as the preferable standard, based on reduced the rates of recurrence, without any significant increase in the postoperative complications found in this group.