Computer simulations have an important role in the study of various social phenomena and concepts. In this paper by doing simulation, various strategies for transfer of experience among agents in artificial society has been proposed and implemented and their effects on some of measures of social welfare like as Gini Coefficient, wealth average and mortality due to starvation have been studied and compared.Research methodology in this paper is simulation based on field and library studies. By using "NetLogo" software, agent-based simulation has been implemented and the required artificial society is created. Also some parameters and measures have been considered in the simulation until thereby the effectiveness of various strategies implementation of experience transfer among agents be measured.Findings and results of simulations indicate that implementation of some of the strategies cause more effective improvement in the measures of social welfare. For example, implementation of the strategy so-called "Training the Least Experienced Agent(s) by All (Other) Agents" compared to other strategies causes better distribution of wealth among agents of the society. Also, strategies of "Collaborative Training", "Training the Least Experienced Agent(s) by All Agents" and "Training the Poorest Agent(s) by All (Other) Agents" have highest effects in better distribution of wealth in the society. Also, in the improvement of wealth average of agents, "Training the Least Experienced Agent(s) by All Agents", and in the reduction of mortality due to starvation, "Collaborative Training", were most effective strategies. The results show that implementation of the strategies in which all agents’ participation involved in the transfer of experience, have most efficiency in improvement of the measures of social welfare.