Speech recognition is the field of research concerning the ability of machines to accept human speech as input and interpret it with the highest probability. There are several techniques for implementing speech recognition models. One of the emerging techniques is to use neural networks with Deep-Learning for speech recognition. Arabic dialect is one of the most widely spoken and least emphasised languages in terms of speech recognition. There has been relatively little research on speech recognition of Arabic dialect compared to other languages. Dialect Arabic has several regional forms and is used for everyday spoken communication in non-formal contexts. With the advent of social media networks, dialectal Arabic is also written. In Morocco, we use Darija dialect as a language of communication. This form of language results in lexical forms of morphological and grammatical differences which results in great difficulty in developing speech recognition. We develop here a deep-learning model based on TensorFlow to recognise this language. This model has been successfully tested in a car driving simulator (Robot Operating System, ROS) with the robot HUSKY with success rates up to 90%[1].