Abstract: In this paper, a quantitative algorithm of human-machine intelligence task allocation is systematically presented for the intelligent system of human-machine cooperation. Basing on the definitions of machine intelligence and human-computer interface intelligence, we introduce a tool of modeling and analyzing system intelligence, namely an intelligent task graph, for the human machine cooperation system. By using the intelligent task graph and its described equations, the entire task to be executed by the intelligent system can be clearly divided into subtasks for human and subtasks for machine, and the human computer interface performance can also be analyzed and optimized. By a calculating example, the validity and importance of the algorithm are verified in the design and analysis of a human-machine cooperation intelligence system.