Income distribution and related efforts to improve it are the most important issues facing government. It is affected by important factors. The present study aims at analyzing tax structure on income distribution. The variables which are used are GINI coefficient as dependent variables, income per capita, education, share of tax on personal income, share of tax on corporate, total income tax and share of tax on goods and services. We apply the panel method to data of 19 countries from 1995 to 2012. Also the model is estimated for Iran. Our hypothesis is that income distribution is influenced by tax structure. The results show that education has no significant effect on income distribution. Kuznets inverse U hypothesis tested in this study and results indicate that Kuznets hypothesis is confirmed in panel model but the country (Iran) has not been approved Kuznets hypothesis. Due to the results of this study, article hypothesis was confirmed. With increasing of personal income tax and corporate income tax share of total tax income, GINI coefficient will be decreased and income distribution will be improved. Also with increasing in share of tax on goods and services, income distribution will be worse and GINI coefficient will be increased. The used cross- section regression confirmed aforementioned results and with increasing in total income tax, income distribution will improve.