This paper presents possibilities of implementation of advanced combustion control concepts in selected Western Balkan thermal power plant, and particularly those based on artificial intelligence as part of primary measures for nitrogen oxide reduction in order to optimise combustion and to increase plant efficiency. Both considered goals comply with environmental quality standards prescribed in large combustion plant directive. Due to specific characterisation of Western Balkan power sector these goals should be reached by low cost and easily implementable solution. Advanced self-learning controller has been developed and the effects of advanced control concept on combustion process have been analysed using artificial neural-network based parameter prediction model. (c) 2013 Elsevier Ltd. All rights reserved.