A guide pair is a core part of the feed system in a machine tool. Its geometric accuracy is attenuated due to wear, thus directly affecting its guiding accuracy and the processing accuracy and quality of a machine tool. In this paper, based on the statistical principle, the influences of the surface wear of guideway on its straightness attenuation was explored and an analytical prediction model for the geometric accuracy attenuation of slide guide was established. The reciprocating wear test of slide guide samples was performed with a test bench to explore the attenuation of guideway surface straightness under various machining conditions: isometric feeding, random length distribution feeding (normal distribution, negatively skewed distribution and positively skewed distribution), and feeding-retracting (with unequal reciprocating speeds). The comparison between experimental results and predicted results showed that the prediction model could well predict the precision attenuation of the guideway under stable working conditions. The experimental results also proved that machining workpieces with the same size or machining workpieces with a larger size as possible was beneficial to extend the precision maintaining life of slide guide. The prediction model suggested the quantitative relationship between the precision attenuation of slide guide and the main factors including material properties, surface topography parameters, working conditions, operation parameters and surface friction properties. With this model, the real-time straightness on the guideway surface can be calculated and the geometric precision maintaining life of slide guide can also be predicted.