The invention discloses a speech recognition method and a speech recognition device, and particularly discloses an age-enhanced speech recognition enhancement method and an age-enhanced speech recognition enhancement device. The age-enhanced speech recognition enhancement device is mainly composed of a speech framing module, a category marking module, an age group coding vector module and a speechrecognition system training module. According to the device, pronunciation category marking needs to be preset for training data framing, a corresponding age group is marked, then an obtained age endcode is added into a speech recognition system training module as a vector, and a speech recognition engine is obtained; after the speech recognition system training module obtains the age vector, speech recognition engine training can be carried out. The invention has the beneficial effects that the age is used as a vector to be added into speech recognition model training through a multi-task collaborative learning method, speech basic information in big data can be shared, and targeted optimization can be performed on crowds of different ages, so that the overall recognition rate of a speech recognition engine is improved.