We present the clustering properties of a sample of infrared (IR) bright dust-obscured galaxies (DOGs). Combining 125 deg^2 of wide and deep optical images obtained with the Hyper Suprime-Cam and all-sky mid-IR (MIR) images taken with Wide-Field Infrared Survey Explorer (WISE), we have discovered 4,367 IR-bright DOGs with (i - [22])_AB > 7.0 and flux at 22 micron > 1.0 mJy. This is the largest sample of IR-bright DOGs in the literature so far. We calculate the angular autocorrelation function (ACF) for a uniform subsample of 1411 DOGs with 3.0 mJy < flux (22 micron) < 5.0 mJy and i_AB < 24.0. The correlation amplitude of IR-bright DOGs is larger than that of IR-faint DOGs, which may reflect a flux-dependence to the DOG clustering. Assuming that the redshift distribution for our DOG sample is Gaussian, with mean and sigma z = 1.99 +/- 0.45 (Case 1) and z = 1.19 +/- 0.30 (Case 2), the correlation length of IR-bright DOGs is r_0 = 13.1 ± 1.3 and 11.3 ± 1.1 h^-1 Mpc for Case 1 and 2, respectively. These correlation lengths indicate that their bias factor is b_DOGs = 6.55 +/- 0.62 and 4.25 +/- 0.40 for Case 1 and 2, respectively. IR-bright DOGs reside in massive dark matter halos with a mass of log[M_h / (h^-1 M_sun)] ∼ 13.76 and 13.86 in Case 1 and 2, respectively, meaning that IR-bright DOGs could be dust obscured progenitors of the most massive galaxy clusters in the present day. Their duty cycle estimated by the ratio of number densities of IR-bright DOGs and dark matter halos in which they reside is at least 0.037 and 0.016 in Case 1 and 2, respectively, which is roughly consistent with that estimated from a hydrodynamical simulation.