Comparative analysis of GSM coverage prediction with measurement results for urban areas using statistical nonparametric mapping Planning, building and optimization process of radio access networks is a complex dynamical activity, which requires a lot of planning effort and time. The final goal is to build a radio network, which consists of minimum infrastructural components, and offers the best coverage and service quality. In order to consider all the possibilities of GSM signal coverage during planning time, and before the installation of equipment on certain location, it is very important to do software simulation of the GSM signal coverage of the observed area. Consequently, it is important to use an optimal predictive model which, using the appropriate digital maps, will offer the most realistic picture of the future radio access network. Taking this into account, subsequent optimization process and re-analysis would be reduced to a minimum. In this paper, comparative analysis of the software prediction GSM coverage area signal with the results obtained by measurements after the implementation of planned radio infrastructure at the proposed base station, will be presented. In order to assess the credibility of the software predicted GSM signal coverage area, measurements of GSM signal in urban area were performed. Final results were verified using statistical nonparametric chi square test. Further, the software prediction relation in regard to the measurement signal after the implementation of system is defined, and potential causes are treated.