Selecting a valid baseline for measuring cardiac autonomic regulation is essential for psychophysiological research. Current guidelines typically recommend eyes-closed resting baselines, preceded by 5 min with eyes open. However, baseline autonomic regulation may vary as a function of both sensory-attentional state (i.e., eye state) and interindividual factors such as physical activity (PA). The present study examined differences between eye-states in autonomic regulation and respiration across a standardized baseline protocol (eyes open [EO1], eyes closed [EC], eyes open [EO2]) and investigated whether these responses are associated with habitual and leisure-time PA. In total, N = 100 university students were recorded during three consecutive 5-min sessions using electrocardiography and impedance cardiography (VU-AMS). Heart rate (HR), root mean square of successive differences (RMSSD), high-frequency power (HF), and respiration rate (estimated via impedance cardiography) were examined. Physical activity was assessed via questionnaires. Mixed-effects models tested condition effects and PA associations, including interactions. Eye state and temporal order significantly modulated cardiac activity, with higher HR and lower vagally mediated HRV during EC and EO2 compared to EO1. Respiration rate differed between EC and subsequent EO2. Low- and moderate-intensity leisure-time PA showed negative associations with vagal indices and interacted with both eyes-open conditions to predict RMSSD. The fixed EO1–EC–EO2 protocol provides reliable baseline estimates consistent with the Vagal Tank Theory and Neurovisceral Integration Model. As the first empirical evaluation of this baseline sequence, the study advances HRV methodology, clarifies context-dependent PA–HRV relationships, and offers normative autonomic data that strengthen experimental and applied psychophysiology.