Enhanced recovery after surgery (ERAS) is an evidence-based multimodal perioperative management strategy. The aim of the present study was to analyze the clinical efficacy of ERAS in elderly patients (> 70 years old) undergoing multi-level posterior lumbar or thoracolumbar instrumented fusion for degenerative diseases. Patients older than 70 years undergoing multi-level lumbar or thoracolumbar instrumented fusion for degenerative disk diseases or spinal stenosis from January 2017 to December 2018 (non-ERAS group) and from January 2020 to December 2021 (ERAS group) were enrolled in present study. Patient-specific and procedure-specific clinical characteristics were collected. Univariate and multivariate regression were performed to determine the risk factors related to length of stay (LOS) and complications. A total of 233 patients were enrolled in this study, 70 in non-ERAS group and 163 in ERAS group. There were comparable baseline characteristics between groups. Further there were no significant differences in 90-day readmission rates and complication rates. However, we observed a significant reduction in LOS (14.89 ± 7.78 days in non-ERAS group versus 11.67 ± 7.26 days in ERAS group, p = 0.002) and overall number of complications (38 in non-ERAS group versus 58 in ERAS group, p = 0.008). Univariate linear regression denoted that operation time (p < 0.001), intraoperative blood loss (p < 0.001), intraoperative blood transfusion (p < 0.001), fusion number ≥ 5 (p < 0.001), spinal surgery including the thoracic spine (p < 0.001), CCI > 2 (p = 0.018), ERAS (p = 0.003) and spinal surgery including lumbar (p = 0.030) were associated with LOS. Furthermore, multivariate linear regression showed that ERAS (p = 0.001), CCI > 2 (p = 0.014), and Fusion number ≥ 5 (p = 0.002) were independent risk factors for LOS. Analogously, univariate logistic regression revealed that longer operation time (p = 0.005), more intraoperative blood loss (p < 0.001), more intraoperative blood transfusion (p = 0.001), fusion number ≥ 5 (p = 0.001), ERAS (p = 0.004) and spinal surgery including thoracic spine (p = 0.002) were related to complications, while implementing ERAS was associated with less complications. Multivariate logistic regression denoted that implementation of ERAS (p = 0.003), Intraoperative blood loss (p = 0.003) and Fusion number ≥ 5 (p = 0.008) were independent risk factors for postoperative complications. In conclusion, the present study reported the first ERAS principles performed in multi-level lumbar or thoracolumbar instrumented fusion for degenerative conditions. Our outcomes shown that the implementation of ERAS in these populations is favorable for reducing LOS and decreasing overall number of complications though the comparable complication rates between two groups. Totally, our ERAS protocols were safe and feasible in these populations.