Abstract: The cycle slip is a peculiar question in the GPS carrier-phase observations.Since the cycle slips that have not been detected are mainly absorbed by the integer ambiguity parameters and receiver position corrections and can lead to the biased estimates for those integer ambiguity parameters,therefore,the detection and repair of cycle slips in carrier-phase data are two major quality-control problems that need to be addressed.There are many methods for detecting and repairing the cycle slips.For example,the polynomial fitting method for detecting the relative big cycle slips is often used.The differences between satellites can detect and repair the cycle slips that related to the satellites.In addition,the smaller cycle slips can be detected and repaired according to the residues obtained after adjustment,and so on.The method called"quasi-accurate detection of gross errors(QUAD)" takes real errors of observations as object to be studied;it can simultaneously detect,locate multiple gross errors in surveying data,estimate their values and accuracies and then correct them.Seeing that the principle of detecting and repairing cycle slips is based on the location of gross error,the QUAD method is introduced to detect and repair the cycle slips in GPS carrier-phase data in this paper.The mathematics model is summarized at first.Then detailed implementation procedure is presented.At last,two numerical examples are given out to illustrate the feasibility and validity of this method.The results show that:① the method is fit for single-frequency as well as for dual-frequency;② if the quasi-accurate observations have been selected correctly,the"hive off"phenomena of real errors are very distinct.Thus using QUAD method for detecting gross errors and cycle slips is very direct and accurate,the gross errors and cycle slips can be detected simultaneously and the estimators of detected gross errors and cycle slips are relatively close to the simulation values as well;③ gross errors and cycle slips can be detected by using very few epochs which is vitally important to the data quality control for Rapid Positioning.