There is a continuous increasing number of subscribers which also demand a high quality of service. This can be met by efficient planning of the radio network, which requires forecasting of the path loss accurately. There are many classical models, but are meant for a specific environment. In this paper experimental measurements in 1.8GHz band were performed for five base stations belonging to dense urban areas of Ghaziabad (Northern India). This experimental data is used in order to compute the experimental path loss. The analysis of forecasting models for path loss under study in this paper are free space model, SUI model, ECC model, Cost 231 Hata model. The comparison between forecasted and experimental path loss shows Cost-231 Hata classical Model to be the closest match having the lowest root mean square error. The forecasting accuracy of Cost 231 Hata model can be enhanced by using a linearly iterative tuning process. The results show a remarkable reduction in root mean square error which approximates to a mean value of 5.38dB, which is quite below the acceptable limit of 6 dB, standard deviation attains a mean value of 5.47dB which is below the acceptable limit of 10dB.