Magnetic hysteresis loops have been measured for polycrystalline MgB2 specimens at temperatures 5 to 30K and under magnetic field up to 1T. Resultant hysteresis width is very narrow, which means the flux pinning is very weak. To such case, a flux creep - flux flow model proposed by Matsushita et al.[1] may be valid. But, the temperature dependence of the irreversibility field was far from their prediction. Direct comparison between the scaling law and the experimental results successfully explains temperature dependence of the J(c), but field dependence of the J(c) is only partly reproduced. It is necessary to introduce an exponential decay with increasing magnetic field proposed by Blatter et al.[ 2] in order to explain the experimental results in high field region.