Based on the point of view that both the macrostructure characters of networks and microstrength of relations have important effects on knowledge flowing, we study the knowledge flowing efficiency of informal knowledge networks with the scalefree character under different interactive policies and levels by computer simulation. The results show that: i) the lower the fuzzy degree of policy adopted, the higher the knowledge flowing efficiency in equilibrium; ii) with the advance of the interactive level, the knowledge flowing efficiency in equilibrium evolves along a Scurve, and an sensitive interval of interactive level emerges. Especially, when based on the degree preferential policy and the network is in equilibrium, the analysis of peak value of average knowledge variance shows that, in terms of average relation strength, a middle value seems to be special.