The invention belongs to the field of radiotherapy and computer computing, and relates to a cloud Monte Carlo state machine architecture method, a storage medium and a cloud Monte Carlo state machinesystem. The cloud Monte Carlo state machine system comprises a main control end, a data processing center and an operation center, wherein the main control end is suitable for dividing a calculation task into a plurality of sub-tasks, distributing the sub-tasks to the operation center for execution, informing the data processing center of the distribution state and task content of the calculationtask executed by the operation center, transmitting input data required by the calculation task to the data processing center, and monitoring the operation center. According to the cloud Monte Carlo state machine architecture system and method provided by the invention, resources can be dynamically allocated, so that the whole Monte Carlo computing system runs in a balanced manner, and the computing efficiency is improved, and cross-platform cross-machine resource scheduling can be realized. Besides, even if the Monte Carlo calculation task is partially completed, for example, a certain operation center cannot complete the calculation task, as long as the completion rate of the calculation task reaches a preset threshold value, task failure cannot be caused.