Heart rate variability (HRV) monitoring is a promising option to estimate the autonomic nervous system regulation responding to exercise. Textiles with embedded sensors recording heartbeat intervals are a simple tool for data collection. The so-called smart shirts offer comfort for daily use and are managed easily. Their measurement accuracy for HRV calculation at rest is promising, but remains questionable during exercise. Therefore, the present study validated the Ambiotex smart shirt using HRV indices (root mean square of successive differences, rel. HF power [high-frequency power percentage of total power] and rel. LF [low-frequency power percentage of total power] power) during exercise. Eighty-three healthy participants (31 ± 6 years; 39 females, 44 males) completed an incremental exercise test on a bicycle ergometer wearing the smart shirt and an electrocardiogram simultaneously. We compared HRV indices of segments at rest (5 min), at warm-up (3 min) and twice at the exercise test (each 5 min). At rest and at warm-up, we observed excellent linear relationship (r > 0.96; R ​​​​​ > 0.94), excellent relative reliability (intraclass correlation coefficient ≥ 0.98; α ≥ 0.98) and acceptable agreement (bias < 10%). During the exercise test, measurement accuracy declined with increasing intensity but remained high (>0.8), although results for partial HRV indices were insufficient. In addition, percentage bias was unacceptable during an exercise test. However, the findings support the validity of the smart shirt for measuring HRV, especially at rest and at warm-up. We suggest using the smart shirt for monitoring HRV indices on a daily basis, but caution should be taken in the interpretation of HRV indices obtained during moderate to vigorous exercise intensities.