A method for recognizing time information from speech data includes: obtaining text information corresponding to speech data; identifying a first time feature contained in the speech data based on the text information; searching within a respective configuration file corresponding to the first time feature to obtain a corresponding time identifier for the first time feature; assigning the corresponding time identifier to a field that is in intermediate time data and that corresponds to the respective configuration file in which the first time feature is located, the intermediate time data comprising multiple fields, and each field being corresponding to a respective configuration file of a plurality of configuration files, each of the plurality of configuration files corresponding a respective category related to time; and obtaining, according to content of one or more fields in the intermediate time data, system time data corresponding to the time information contained in the speech data.