Quantitative analysis of time scale effects is conducive to further understanding of vegetation water and soil conservation mechanism. Based on the observation data of the grass covered and bare soil (control) experimental plots located in Hetian Town, Changting County of Fujian Province from 2007 to 2010, the characteristics of 4 parameters (precipitation, vegetation, RE and SE) were analyzed at precipitation event, month, season, and annual scales, and then the linear regression models were established to describe the relationships between RE(SE) and its influencing factors of precipitation and vegetation. RE (SE) means the ratio of runoff depth (soil loss) of grass covered plot to that of the control plot. Results show that these 4 parameters presented different magnitude and variation on different time scales. RE and SE were relatively stable either within or among different time scales due to their ratios reducing the influence of other factors. The coupling of precipitation and vegetation led to better water conservation effect at lower RE (< 0.3) at precipitation event scale as well as at season scale, while the water conservation effect was dominated by precipitation at slightly higher (0.3-0.4) and higher (>0.7) REs at precipitation event scale as well as at annual scale (R2 > 0.78). For the soil conservation effect, precipitation or/and vegetation was/were the dominated influence factor(s) at precipitation event and annual scales, and the grass LAI could basically describe the positive conservation effect (SE<1，R2>0.55), while the maximum 30 min intensity (I30) could describe the negative conservation effect more accurately (SE>1, R2>0.79). More uncertainties (R2≈0.4) exist in the models of both RE and SE at two moderate time scales (month and season). Consequently, factors influencing water and soil conservation effect of grass present different variation and coupling characteristics on different time scales, indicating the importance of time scale at the study on water and soil conservation.