Fuel injector is the key part of automobile engine and its spray characteristic is an important factor influencing the performance of its work. The spray angle of fuel injectors is one of the most important parameters in the spray characteristic, and needed to measure in the detection of fuel injector performance. General method of measurement is collecting multiple injections of fuel by patternation, but it is poor accuracy and efficiency. And for gasoline-direct injection injectors, the mass lost to vapor can be very substantial, so lead to measurement errors for unable to collect fuel spray completely. Using machine vision technology can detect the spray angle, and using high-speed cameras to capture the injectors spray image, whose shadow can be eliminated by the microsecond exposures for ensuring the quality of spray image. Then the spray angle gets image preprocessing, image segmentation, and calculation. Experimental results show that this method can detect fuel injector spray angles quickly and accurately, and successfully applied to the injector measuring test-bed developed by research group.