To study the absorption characteristics of rhodopsin, a dim-light photoreceptor, in chub mackerel (Scomber japonicus) and the relationship between light wavelengths on the photoresponse, the rod opsin gene was cloned into an expression vector, pMT4. Recombinant opsin was transiently expressed in COS-1 cells and reconstituted with 11-cis-retinal. Cells containing the regenerated rhodopsin were solubilized and subjected to UV/Vis spectroscopic analysis in the dark and upon illumination. Difference spectra from the lysates indicated an absorption maximum of mackerel rhodopsin around 500 nm. Four types of light-emitting diode (LED) modules with different wavelengths (red, peak 627 nm; cyan, 505 nm; blue, 442 nm; white, 447 + 560 nm) were constructed to examine their effects on the photoresponse in chub mackerel. Behavioral responses of the mackerels, including speed and frequencies acclimated in the dark and upon LED illumination, were analyzed using an underwater acoustic camera. Compared to an average speed of 22.25 ± 1.57 cm/s of mackerel movement in the dark, speed increased to 22.97 ± 0.29, 24.66 ± 1.06, 26.28 ± 2.28, and 25.19 ± 1.91 cm/s upon exposure to red, blue, cyan, and white LEDs, respectively. There were increases of 103.48 ± 1.58, 109.37 ± 5.29, 118.48 ± 10.82, and 109.43 ± 3.92 %, respectively, in the relative speed of the fishes upon illumination with red, blue, cyan, and white LEDs compared with that in the dark (set at 100 %). Similar rate of wavelength-dependent responses was observed in a frequency analysis. These results indicate that an LED emitting a peak wavelength close to an absorption maximum of rhodopsin is more effective at eliciting a response to light.