When developing driver assistance systems based on machine vision, many environmental factors should be considered. This leads to complex algorithms and codes. Moreover, in early developing stage, the system can hardly be validated effectively, since it is impractical to do real vehicle tests because of security reasons. To solve this problem, a rapid development and test platform for driver assistance systems based on machine vision is designed by using NI EVS and PXI. EVS is suitable for fast developing of image processing and detection algorithms. It can connect multiple cameras and realize synchronous acquisition. With LabVIEW and VDM module, the developing process of detection algorithm and its product is greatly simplified and improved. The real time platform based on PXI and the virtual reality software CarMaker are integrated by VeriStand. Thereby the developing platform is set up, which can simulate effective virtual test scenarios for comprehensive, effective and repeatable tests of machine vision systems.