The purpose of this study was to identify whether the preoperative hemoglobin to albumin ratio (HAR) could predict the prognosis of patients who underwent colorectal cancer (CRC) radical resection. This study enrolled 4018 consecutive CRC patients, calculating HAR as the hemoglobin count divided by albumin count. Patients were divided into the high and low HAR groups based on a cut-off value (0.36). Baseline information and short-term outcomes were compared between the two groups. Logistic and Cox regression analyses were conducted to determine whether HAR was an independent risk factor for CRC. A total of 4018 patients were divided into the high HAR group (3295) and the low HAR group (723). It was found that the high HAR group had more females (P < 0.01), less BMI (P = 0.027), less smoking (P < 0.01), less drinking (P < 0.01), less T2DM (P = 0.027), lower albumin (P < 0.01), higher hemoglobin (P < 0.01) and more rectal cancer (P = 0.026). We found that HAR was an independent risk factor for overall complications (P = 0.012, OR = 1.279, 95% CI 1.055-1.550). Moreover, we found that HAR was an independent risk factor for overall survival (OS) (P = 0.012, HR = 1.300, 95% CI 1.059-1.597) and disease-free survival (DFS) (P = 0.030, HR = 1.231, 95% CI 1.021-1.484). We found that the low HAR group had worse OS in stage III (P = 0.012) CRC than the high HAR group. In terms of DFS, the low HAR group also had worse DFS in stage III (P = 0.01) CRC than the high HAR group. HAR was an independent predictive factor for the prognosis of CRC. Therefore, surgeons should pay attention to hemoglobin and albumin values before surgery.