Embodiments of the invention disclose a method and a device for detecting cheating, and relate to the technical field of e-commerce. The method and the device can reduce audit cost, and further improves depth of automatic investigation. The method comprises: counting sample data, to obtain index parameters of each sample, dimension of the index parameters of one sample including total click rate of one sample, click number of people, click rate in per hour, a sample click region, and click rate of a search term corresponding to the sample; according to the index parameters, generating feature parameters of each sample, the dimension of the feature parameters of one sample including a click rate dispersion coefficient, a dispersion coefficient of click rate in per hour, terrain concentration ratio, a query term diversity index, and a per capita click rate; according to the feature parameters of each sample, establishing a detection model, and according to the detection model, determining whether to-be-tested commodity data is abnormal or not, the dimension of the detection model including the dimensions of the feature parameters of each sample. The method and the device are suitable to be used for automatically checking cheating commodities and users.