The main goal of this paper is to define the determinants of the regional disparities and to establish the factors of regional convergence and their contribution to reducing regional inequalities pursuant to integration and development groups in the EU, and all in purpose of researching the effects of mentioned inequalities on the increase of economic and social cohesion and the development of internal market. The analysis begins by reviewing theoretical postulations on the issues of national and regional development since the mid-20th century till today, and continues with the explanation of types and models of convergence with the particular emphasis on beta convergence and sigma convergence and the modalities of their measurement. Starting from the premise that regional unevenness is a major limitation of proportioned economic development, an analysis of regional unevenness in EU-15, EU-10 and EU-27 was made, and the results have confirmed the existence of strong and complex regional inequalities. In old member states (EU-15) there are differences between the developed north and less developed south, while in new member states (EU-10) there are differences between metropolitan and western regions on one side and the eastern regions on the other side. European Union (EU-27), along with the development of the metropolitan regions and core regions, is characterized by strong regional unevenness between the old and new member states as well as the general underdevelopment of peripheral areas. The index values show that after the accession of new member states there have been increase of the regional differences and decrease of internal cohesion, which resulted in the increase of vulnerability to asymmetric shocks. Comparative analysis of regional policies in old and new member states identified the differences in investments and priority goals. Analysis performed on the interdependence of GDP level and the components of the research and knowledge has shown that regional GDP does not depend on investments in R&D. Among all analyzed components of research and knowledge, the strongest influence on GDP has the number of people employed in science and technology. The analysis of the interdependence of the components of research and knowledge and GDP growth rate has not found firm and unambiguous connection between mentioned components. This proves that components of research and knowledge are not key determinants for regional growth, but rather unique components of growth and development can be determined at each region level. Analysis of regional competitiveness shows that the most competitive regions are core regions and northern regions, and the least competitive are southern regions and regions of eastern periphery. Bad results and widely set goals of regional policies open the controversies between the principle of solidarity and efficiency and identify the necessity of strong coordination at all levels of management. Although the structural funds, common policies and the approach to the unique market have obtained certain positive effects on the regional GDP and employment, it is obvious that the comprehensive progress concerning the employment and internal cohesion is absent. In the following step a model of regional convergence was formulated which was then tested through the econometric techniques according to integration and development groups. The results of the testing in all observed integration groups of regions show that the industrial and construction investments, employees in industry, employees in services, compensations of employees and productivity are main determinants of regional convergence. Although all observed groups of regions realize beta-convergence, the regions with the above-average GDP can faster achieve balance and in that way the period for achieving balance for the poorer regions elongates. The results of testing the convergence model in development groups show that investments, private household income, gross value added in industry and productivity are main determinants of convergence. On the contrary to many theoretical models and propositions, results show that regional convergence is not firmly leaned to the components of research and knowledge, but is rather based on many different factors that are firmly and unambiguously embedded in regional economic systems. After testing the beta convergence, the establishing of sigma convergence as an indicator of regional differences trends was approached. EU-15 regions record slight sigma convergence to year 2008 when due to effects of economic crisis a slight divergence movement is discerned, while EU-10 regions after the accession do not realize clear either convergence or divergence trends. According to development groups sigma convergence shows convergence trends in group of regions of goal 1, while for other groups of region there are divergence trends. New member states (EU-10) converge to countries of hard core (EU-15) at national level, but at regional level processes of divergence between metropolitan and other regions are more expressed. Conducted research shows that, due to constant increase of stress on financial funds and common policies, EU should rely more on realization of internal market and less on current programs and common policies.