Fetal well-being during labor is usually assessed by visual analysis of a fetal heart rate (FHR) tracing. Our primary objective was to evaluate the ability of automated heart rate variability (HRV) analysis methods, including our new fetal stress index (FSI), to predict neonatal acidosis. 552 intrapartum recordings were analyzed. The analysis occurred in the last 90 min before birth and was conducted during two 5-min intervals: (i) a stable period of FHR and (ii) the period corresponding to the maximum FSI value. For each period, we computed the mean FHR, FSI, short-term variability (STV), and long-term variability (LTV). Visual FHR interpretation was performed using the FIGO classification. The population was separated into two groups: (i) an acidotic group with an arterial pH at birth ≤ 7.10 and a control group. Prediction of a neonatal pH ≤ 7.10 was assessed by computing the receiver-operating characteristic area under the curve (AUC). FHR, FSI, STV, and LTV did not differ significantly between groups during the stable period. During the FSI max peak period, LTV and STV correlated significantly in the acidotic group (– 5.85 ± 2.19, p = 0.010 and – 0.62 ± 0.29, p = 0.037, respectively). The AUC values were 0.569 for FIGO classification, 0.595 for STV, and 0.622 for LTV. The multivariate model (FIGO, FSI, FC, STV, LTV) had the greatest accuracy for predicting acidosis (AUC = 0.719). FSI was not predictive of neonatal acidosis probably because of the low quality of the FHR signal in cardiotocography. When used separately, HRV indexes and visual FHR analysis were poor predictors of neonatal acidosis. Including all indexes in a multivariate model increased the predictive ability.