The combined index of hemoglobin, albumin, lymphocyte, and platelet (HALP) can reflect systemic inflammation and nutritional status simultaneously, with some evidence revealing its prognostic value for some tumors. However, the effect of HALP on recurrence-free survival (RFS) in patients of gastrointestinal stromal tumors (GISTs) has not been reported. Therefore, the present study aimed to investigate the prognostic value of HALP in GIST patients. Methods Data from 591 untreated patients who underwent R0 resection for primary and localized GISTs at West China Hospital between December 2008 and December 2016 were included. Clinicopathological data, preoperative albumin, blood routine information, postoperative treatment, and recurrence status were recorded. To eliminate the baseline inequivalence, the propensity scores matching (PSM) method was introduced. The relationship between RFS and preoperative HALP was investigated. Results The optimal cutoff value for the HALP was determined by the x-Tile analysis at 31.5. HALP was significantly associated with tumor site, tumor size, mitosis, Ki67, NIH risk category and adjuvant therapy (all P<0.001). Before PSM, GIST patients with an increased HALP had a significantly poor RFS (P < 0.001), and low HALP was an independent risk factor for poor RFS (HR=0.0551, 95% CI: 0.313 - 0.968, P=0.038). In NIH high-risk GIST patients, GIST patients with low HALP had a worse RFS than patients with high HALP (P<0.05). After PSM, 188 pairs of GIST patients were identified, GIST patients with an increased HALP still had a significantly poor RFS after PSM (P<0.001), and low HALP was still an independent risk factor for poor RFS (HR=0.585, 95% CI: 0.316 - 0.972, P=0.042). Conclusions HALP had a statistically significant correlation with postoperative pathology and postoperative treatment. Furthermore, HALP has a strong ability to predict the RFS in GIST patients with radical resection.