Nondefault inconsistent hourly precipitation data are an abnormal status in automatic weather station (AWS) observation, which can be often met as hourly precipitation data are transmitted and recorded in 2 sources. Three groups of related instances are listed first, and the direct reasons that deeply hidden are found out manually. To solve this problem, Inter VerificationSequence Alignment (IVSA) model in smaller time scale is raised in this article. When nondefault inconsistent data from same station appears at the same time, verification with smaller time scale data (minute precipitation) is made respectively. If both data cannot be proved wrong with inner verification step, then unit error possibility is added into sequence alignment method. Correlation credibility is calculated and more reliable data can be selected accordingly. After then, monthly data in May 2012 (1360 pairs of instances) are used to train the parameters, and the data (4017 pairs of instances) from 1 June to 31 July 2012 are input to verify the efficiency of applying IVSA in real time data environment, getting an accuracy of 99.65%. It is concluded that IVSA model can eliminate nondefault inconsistence in hourly precipitation data under running rules.