High-precision, rapid and non-contact geometry parameter measurements of plate glass parts has become a main problem in the field of testing-related production and processing and also an important applied direction for laser spectroscopy. Accurately detect the geometric parameters of plate glass parts can not only improve the processing technology and the precision of assembly, but also bring about sub-file management according to the parameters. This paper presents a novel multi-parameter measurement method based on the laser and vision image processing technology, which can be used to measure parameters of miniature quartz plate-sensitive glass part accurately. The testing system consisting of self-adaption coaxial visual detection unit and laser vision thickness inspection unit was designed. A constant power drive control system was set up to ensure the laser diode (LD) can provide a stable light source for the testing system. A modified sub-pixel edge position algorithm of conic features was proposed to implement the sub-pixel image processing, and the parameters can be extracted. According to the data detected from curve edge points and a new-defined error function, which was minimized, the parameters of miniature quartz plate-sensitive glass part can be calculated. The experiments show that the average deviation of the measurement results is less than 2 μm, and the method has good stability and high accuracy measurement, which can meet the precision requirement of parameter measurement for miniature plate glass part.