Husbandry training (HT) shapes specific actions of animals to facilitate healthcare. The challenge of HT is that the learning of the specific actions requires zoo-keepers to grasp the stress states. Here we suggest non-wearable pulse rate measurement system, as pulse rate indicates stress state. The system uses laser Doppler flowmetry (LDF) with an algorithm to eliminate body motion artifacts. In the algorithm, the cutoff frequency is an important parameter. For the masked palm civet (Parguma larvata) during HT, sensor signal was recorded. We applied 100 Hz, 12 kHz, and 15 kHz cutoff frequencies to the same sensor signal and calculated blood flow signals. When pulse waveforms were confirmed, blood flow signals of 100 Hz and 12 kHz indicate a strong peak at 3.2 Hz (192 beats/min) in the frequency analysis. When pulse waveforms were disturbed, blood flow signal of 15 kHz indicates the strong peak at 3.1 Hz (186 beats/min).