An electronic apparatus and a method of speech recognition thereof are disclosed. According to the method of speech recognition of the electronic apparatus, the method includes receiving a speech of a speaker, extracting phonemic characteristics for recognizing a speech and voice print characteristics for registering the speaker by analyzing the received speech of the speaker, and in response to the speech of the speaker corresponding a registered trigger word or phrase, based on the extracted phonemic characteristics, changing an execution mode to a speech recognition mode of the electronic apparatus and registering the extracted voice print characteristics as voice print characteristics of the speaker who spoke the speech.