Unpredictable rainstorms can pose greater risks of soil erosion on newly formed unprotected railway batters. Bio-engineering approach consisting of revegetation can be an effective tool for soil erosion control. Perennial species, such as buffel grass, are preferred for durability of vegetation cover in the arid and semi-arid tropics. However, it takes relatively longer time to establish buffel compared with the annual species such as Japanese millet (JM). A replicated plot-scale field trials consisting of different proportions of grass seed mixtures of buffel and JM was conducted on railway batters slope in the semi-arid tropics of central Queensland, Australia. Supplementary automated drip irrigation system was set up to aid seed germination and early establishment of the grass cover. JM alone or JM dominated seed mixtures significantly reduced soil erosion during the earlier stages. For example, during the second sampling interval, which occurred 63 days after seeding, JM alone reduced soil erosion by 50% (20.2 vs 10.21 t/ha) compared with the sole buffel. This trendbegan to shift as the growth of buffel took up at the latter stages. By the end of the 11 months, erosion from all seeded plots was not significantly different from each other with over 90% soil loss reduction compared to the bare scenario as a result of over 60% grass cover. Results suggested that a monoculture of JM, or JM dominated seed mixture with buffel grass, could be an effective bioengineering approach for revegetation of railway batters for protection against soil erosion. Quick establishment due to faster seed germination, seedling establishments and growth producing sizable canopy cover developed by JM in early stage of batter could reduce soil erosioncompared to sole crop of slow growing perennial species, buffel. The established vegetation, self regenerated, and responded dynamically and naturally to changing conditions of the railway batter within a year, after which JM started to die back.