Abstract： Amount of historical anecdotes suggest that the heuristic prototype is an effective method to get insight. The prototype heuristic theory suppose that insight obtain when people activate a prototypical example in nature and apply it to the problem at hand. In the real life, people often encounter a problem and then a heuristic prototype help them solve this problem. Some previous studies used the quartering problem and demonstrated that the automatic activation of prototype was the key process of prototype heuristic. However, this relationship was needed to test with real-life scientific innovation problems. Moreover, the mechanism of prototype activation also needed to explore further. Thus, present study used two experiments to test the cognitive mechanism of the prototype activation with real-life scientific innovation materials. In experiment 1, forty-eight participants were recruited to complete the prototype heuristic task. There were two stages in this task. In the first stage, participants needed to learn ten scientific problems that were randomly presented. They should press some buttons to indicate whether they have some knowledge of these problems. In the second phase, ten prototypes were presented and participants should first write which scientific problem learnt in the first stages could be solved by the current prototype. Then, participants needed to write how to solve the previous scientific problem with current prototype. The dependent variables were the prototype activation and the accuracy of the problem solving. In experiment 2, forty-seven participants completed the prototype heuristic task. The procedure was the same with experiment 1 and the prototype activation was the dependent variable. In experiment 1, a regression analysis was conducted. In this analysis, the problem activation was the independent variable and the accuracy of the problem solving was the dependent variable. The result of the regression analysis was R2 = 0.893, p < 0.001. This result suggested that the problem activation was the key process of problem solving. The results of experiment 2 showed that retrieval of the key words of the scientific problem significantly influenced the association of the problem and the prototype, t(44)= 27.677, p < 0.001. Moreover, match of the key words of the prototype and scientific problem significantly influenced the activation of the problem, t (44) = 8.744, p < 0.001. In addition, semantic similarity between the key words of the scientific problem and the prototype was significantly correlated with the problem activation, r = 0.202, p < 0.05. The results also showed that the difficulty of the prototype significantly influenced the problem activation, F(2,132)= 96.73, p < 0.001, h2p = 0.594. In summary, the results suggest that problem activation was the key process in the real-life problem solving enhanced by heuristic prototype. The semantic similarity between the feature function of the prototype and the required function of the problem is the mechanism of the problem activation.