In this paper, the modular role player game-distance-learning system based on voice recognition for language learning in cloud is proposed. Each designed modular game content can be embedded into the learning game platform in cloud. By connecting to the external recognition resource, the pronouncing of learners can be recognized. Only when the pronouncing of the learner can be recognized as the keyword, the further game will be started. By designing different scenarios, different vocabularies or conversations can be added into the RPG game. The information about users in the individual game learning system can be stored in the cloud database. Then users can connect to the learning-game system via the Internet for continuous learning. By using the cloud platform, the users can connect to learning system anytime and anywhere. Furthermore, the students can learn advanced courses themselves by playing the advanced scenarios of the game-learning system without the assistance of a teacher.