Objective: Investigate whether patterns of heart rate variability (indexed via respiratory sinus arrhythmia) and visual attention at 12 to 18 months of age predict elevated ADHD symptoms, autism, or neither during the preschool period. Method: Ninety infants 12 to 18 months of age (M = 17.27, SD = 1.93; 36 females; 82.2% non-Hispanic) participated in a split-screen eye-tracking task of dynamic social and non-social moving objects. Respiratory sinus arrhythmia was derived from heart rate data collected at baseline and during the task condition. Between 24 and 65 months of age (M = 38.22, SD = 11.14), participants were evaluated and classified into one of three outcome groups: ADHD Concerns (n = 21), Autism (n = 12), or Comparison (i.e., non-Autism/non-ADHD Concerns; n = 57). Results: The ADHD Concerns group exhibited significantly less whole-screen looking time (t(76) = -2.98, p = 0.004, d = 0.82) and spent a significantly lower proportion of time attending to the social portion of the stimulus (t(76) = -2.53, p = 0.01, d = 0.67) than the Comparison group. Respiratory sinus arrhythmia reactivity moderated the association between proportion of time spent looking at the social portion of the stimulus in infancy and ADHD symptoms during the preschool period (b = 0.004, 95% CI [0.0001, 0.01], t(89) = 2.11, p = 0.04), such that greater quantity/intensity of ADHD symptoms was associated with a smaller proportion of look time to the social portion of the stimulus for infants engaging in HRV withdrawal, but not HRV augmentation. Hypotheses focused on autism were not supported. Conclusions: Infants demonstrate distinct patterns of visual attention predictive of elevated ADHD symptoms in the preschool period. Heart rate variability may also demonstrate predictive utility in the context of early ADHD when examined in relation to social attention, but not independently.