The cloud robotic technology makes multiple robots share resources and collaborate with each other more flexibly. Although this concept has been widely accepted by academia since it was put forward, it still faces many challenges in engineering. The main problems are that the functions of industrial robotic systems become more and more complex, and the programming language and computing environment of multiple vendor robots are significantly different, accordingly the cooperative control of different devices and the online testing become extremely difficult. Therefore, this paper proposes a new control platform for cloud robots, called IAPcloud, which has significant advantages in solving the problem of collaborative control among heterogeneous robots and their auxiliary devices, declining the programming difficulties of cloud robotic control systems and shortening the development and deployment cycle of the application. Finally, we verify the scientific and effectiveness of the IAPcloud platform by three cases, including the image recognition and tracking, the human–machine gobang game, and the online dynamic reconfiguration of control algorithms.