A method and a device for voice controlling are provided, which are used to solve the problem of low mission success rate of voice controlling in the prior art. The method includes: classifying the stored recognition information used for recognizing a voice to obtain each syntactic packet corresponding to the recognition information (10); receiving an inputted voice signal, and performing respectively a voice recognition processing on the received voice signal by using the each syntactic packet (20) obtained in turn, and performing a corresponding control processing according to the voice recognition result of the voice signal produced by each syntactic packet (30).