Structural best management practices (BMPs) are often used to mitigate the impact of storm water runoff on receiving waters. Vegetative filter strips (VFS) are an example of a structural BMP that has been used to treat storm water and highway runoff. Physical factors affecting the performance of VFS include pollutant characteristics, vegetation composition and density, soil properties, and the physical dimensions of the filter strip. In this study, field-suspended sediment data were collected from an experimental VFS treating highway runoff in eastern North Carolina. Field data were used to test the design concepts of the VFS treatment train and to validate a simulation model for evaluating the impact of these physical factors on sediment removal as a function of filter strip length. It was concluded that the experimental filter strip was effective in removing more than 85% of the incoming total suspended sediment (TSS). Simulation results support field observations that a 10-m or longer filter strip can retain most of the medium and large particles (> 8 microm) transported in runoff. Simulations also indicate infiltration loss is largely responsible for the retention of small-size sediment particles (< 8 microm). Saturated hydraulic conductivity and initial water contents have little effects on TSS removal. The condition of vegetative coverage, in particular vegetation density, is another factor affecting the performance of filter strip.