Purpose: We aimed to retrospectively analyze the clinical features, laboratory and imaging results, and predictors of poor prognosis for patients with an initial diagnosis of autoimmune encephalitis (AE) at the Affiliated Hospital of Zunyi Medical University.Methods: Fifty patients with an initial diagnosis of AE who were admitted to our hospital from May 2014 to May 2018 were enrolled retrospectively. Clinical characteristics and experimental test data, including the neutrophil-to-lymphocyte ratio (NLR), were collected from medical records within 24 h of admission. Independent prognostic factors were determined by multivariate logistic regression analysis. A good or poor prognosis for patients was defined based on the modified Rankin Scale (mRS). The correlation between the immunotherapy latency and prognostic mRS score was determined using the Spearman rank correlation test.Results: Univariate analysis indicated that increased NLR (P = 0.001), decreased lymphocyte counts (P = 0.001), low serum albumin (P = 0.017), consciousness disorders (P = 0.001), epileptic seizures (P = 0.007), extrapyramidal symptoms (P = 0.042), abnormal electroencephalogram (EEG) findings (P = 0.001), abnormal brain magnetic resonance imaging (MRI) findings (P = 0.003), and pulmonary infection complications (P = 0.000) were associated with the poor prognosis of AE. Multivariate logistic regression analysis showed that NLR (odds ratio [OR] 2.169, 95% confidence interval [CI] 1.029–4.570; P < 0.05) was an independent risk factor for predicting the poor prognosis of AE. NLR > 4.45 was suggested as the cut-off threshold for predicting the adverse outcomes of AE. In addition, we revealed that there was a positive correlation between immunotherapy latency and mRS score (rs = 0.535, P < 0.05).Conclusions: NLR may have predictive value for the poor outcomes of AE. Early initiation of immunotherapy is associated with a good prognosis.