First stars can only form in structures that are suitably dense , which can be parametrized by the minimum dark matter halo mass M _ { min } . M _ { min } must plays an important role in star formation . The connection of long gamma-ray bursts ( LGRBs ) with the collapse of massive stars has provided a good opportunity for probing star formation in dark matter halos . We place some constraints on M _ { min } using the latest Swift LGRB data . We conservatively consider that LGRB rate is proportional to the cosmic star formation rate ( CSFR ) and an additional evolution parametrized as ( 1 + z ) ^ { \alpha } , where the CSFR model as a function of M _ { min } . Using the \chi ^ { 2 } statistic , the contour constraints on the M _ { min } – \alpha plane show that at the 1 \sigma confidence level , we have M _ { min } < 10 ^ { 10.5 } M _ { \odot } from 118 LGRBs with redshift z < 4 and luminosity L _ { iso } > 1.8 \times 10 ^ { 51 } erg s ^ { -1 } . We also find that adding 12 high- z ( 4 < z < 5 ) LGRBs ( consisting of 104 LGRBs with z < 5 and L _ { iso } > 3.1 \times 10 ^ { 51 } erg s ^ { -1 } ) could result in much tighter constraints on M _ { min } , for which , 10 ^ { 7.7 } M _ { \odot } < M _ { min } < 10 ^ { 11.6 } M _ { \odot } ( 1 \sigma ) . Through Monte Carlo simulations , we estimate that future five years of Sino-French spacebased multiband astronomical variable objects monitor ( SVOM ) observations would tighten these constraints to 10 ^ { 9.7 } M _ { \odot } < M _ { min } < 10 ^ { 11.3 } M _ { \odot } . The strong constraints on M _ { min } indicate that LGRBs are a new promising tool for investigating star formation in dark matter halos .