Objective: To assess the clinical efficacy and safety of imatinib treatment for advanced and high-risk gastrointestinal stromal tumor (GIST). Methods: The clinical data of 173 GIST patients treated between January 2011 and June 2016 were reviewed. Of the patients, 73 cases had advanced GIST, and 100 cases had high-risk GIST. The outcomes were compared between patients with and without imatinib treatment in these two categories of patients, respectively. Results: The 73 patients with advanced GIST were followed up for 31 (6–66) months, and of them, the Cox regression analysis showed that the overall survival (OS) in patients with imatinib treatment was significantly prolonged compared with those without imatinib treatment (1-year OS: 100.0% vs. 78.6%, 2-year OS: 93.1% vs. 26.1%; HR=0.040, 95% CI=0.011–0.152, P=0.000). Follow-up was conducted for 45 (6–73) months in the 100 patients with high-risk GIST, of whom, the subgroup analysis showed that the recurrence free survival (RFS) in patients receiving postoperative 1-year imatinib treatment was significantly longer than that in those without imatinib treatment (3-year RFS: 66.7% vs. 38.5%; HR=0.341, 95% CI=0.134–0.868, P=0.024) and moreover, the RFS in patients receiving postoperative 2-year imatinib treatment was significantly longer than those receiving 1-year imatinib treatment (1-year RFS: 100.0% vs. 100.0%, 2-year RFS: 100.0% vs. 88.9%, 3-year RFS: 91.7% vs. 66.7%; HR=0.108, 95% CI=0.015–0.778, P=0.027); the 3-year RFS was 100.0% in the 5 high-risk patients who received postoperative 3-year imatinib treatment. The common adverse effects from imatinib treatment were edema, leukopenia and gastrointestinal disorders, and most of them were grade 1 to 2 in severity. Conclusion: Imatinib has favourable safety in treatment of advanced and high-risk GIST, and it can effectively improve the survival of the patients. For those high-risk patients, at least 3 years’ postoperative imatinib administration is recommended, but whether to extend the 3-year time limit of drug administration needs to be determined by further clinical studies.