Anti-programmed cell death receptor (PD)-1 antibody treatment results in better prognosis than standard chemotherapy in patients with non-small cell lung cancer (NSCLC), especially those with high PD-ligand 1 (PD-L1) expression. However, several studies have reported a lack of antitumor effect of PD-1 antibody, even in patients with high PD-L1 expression. Therefore, reliable predictors of treatment response are urgently needed. The albumin-globulin ratio (AGR) is associated with prognosis in several cancers. We aimed to determine whether AGR is a predictive biomarker of anti-PD-1 antibody response in patients with NSCLC. Seventy-four NSCLC patients treated with anti-PD-1 antibody were retrospectively enrolled. Patients with driver mutations were excluded. The mean AGR was significantly higher in the disease control (DC) group than in the progressive disease (PD) group (p < 0.001). Receiver operating characteristic curve analysis revealed an AGR cutoff value for dividing patients into the DC or PD groups of 1.17. Multivariate logistic regression analysis showed that a high AGR (≥1.17, cutoff value) was an independent predictor of DC (p = 0.001). Progression-free survival (PFS) and overall survival (OS) were significantly longer in the high-AGR group than in the low-AGR group (p = 0.008, p = 0.002, respectively). Multivariate Cox regression analysis of PFS and OS showed that high AGR was an independent prognostic factor (p = 0.020, p < 0.001, respectively). Pretreatment serum AGR may be a useful predictor for DC and prognostic factor of anti-PD-1 antibody in patients with NSCLC. The clinical utility of AGR still needs to be confirmed in a prospective analysis.