A nation's ability to develop technologically, socially, politically, and economically depends on a reliable supply of electricity. It is impossible to imagine modernization and development in any field without a reliable electric power source. Due to the frequent and prolonged unplanned power outages in Ethiopia, there is a severe problem with power interruptions. Because of this, utilities must work hard to meet both regulatory requirements and consumer reliability requirements while keeping cost to a minimum. In this thesis, the performance of 15kV Sebeta electric power distribution system is assessed, and reliability improvement for the future is predicted while taking distribution system reconfiguration into consideration (optimal number and location of switches). For predictive reliability improvement, MATLAB modeled fuzzy logic based optimization technique has been applied, which delivers optimal number and placement of switches on the selected feeder of the distribution system. Among the 13 feeders of the distribution system under case study, feeder-3 is selected as a test feeder as it has highest frequency and duration of interruption and also its cost of interruption is very high compared to other feeders. In general, this thesis work succeeds that reliability of the system has been improved significantly by applying fuzzy logic based optimal placement of switches. From Electrical Transient Analysis Program (ETAP) simulation result of the proposed method, System Average Interruption Frequency Index (SAIFI) can be improved from 208.596 f/yr to 53.4075 f/yr which is 74.39%, and System Average Interruption Duration Index (SAIDI) can be improved from 229.5703 hr/yr to 50.1439 hr/yr which is 78.15%. Key Words: ETAP software, Five Whys Analysis, Fuzzy Logic Optimization, Reliability Benchmarks, Root Cause Analysis