Universal Mobile Telecommunications System (UMTS) with wide-band code division multiple access technology (WCDMA) is a major mobile communication system deployed in Ethiopia. WCDMA is being widely deployed in urban areas of the country. Along its deployment and expansion, it is necessary to develop and maintain a good quality of service (QoS) for mobile subscribers. To realize a certain network QoS, one of the prior works to be done is proper capacity planning before deployment. A discrepancy between the installed capacity of an operator and capacity demand of its customers results in inefficiency, either in under-utilized resources or unsatisfied customer. Improper capacity planning can lead to high capital cost or loss of revenue due to unsuccessful calls and low data rate (speed). So, working on capacity analysis needs prudent decisions.rnAs a network provider, when considering deployment of new networks or expansion of existing networks, knowing the required capacity is a core part of the overall planning process. Typically, forecasted traffic parameters play a major role in determining the user traffic model and dimensioning the total number of users in a cell. Capacity demand may vary based on input parameters, such as busy hour traffic per user, downlink and uplink ratio, and busy hour traffic ratio.rnThe goal of this thesis is forecast required voice and data capacity of ethio telecom UMTS network using artificial neural network (ANN) model. The forecasting result could be used to minimize discrepancy during planning. Taking Addis Ababa city as a research focus area, real time UMTS traffic measurements were obtained for a period of 14 consecutive months from the Performance Reporting System (PRS). Traffic model forecasting is done by using ANN model. Real time traffic measurement counters are obtained based on the actual voice call traffic and total throughput of data service traffic values. The variation of each input assumption is analyzed with reference to the deployed capacity planning result. The variation due to each input assumption is analyzed with UMTS traffic model cell load analyzer tool which is developed for this study.rnUMTS Traffic Model Using ANN: The Case of Addis Ababa, Ethiopia iirnUsing the operator forecasted total number of subscribersâ€™ result data, the result of this study shows that the holding capacity of the network will get efficient if the number of cells remain the same as of the deployed result for the next five years.rnKey Words: UMTS; Traffic model; Capacity Planning; Artificial Neural Network model.