The unreliability of the electrical grid has been a major challenge in Jigjiga City. The existing substation has encountered frequent power interruption problems. The interruptions are caused mainly by the short circuit (SC) and earth fault (EF).There are also planned outages for operation and maintenance and the interrupted electric power distribution reduce power consumption ,affect daily activity and drags the modern life style. It impacts societal development and individual income. Basically, a stable and reliable electric power supply system is an inevitable pre-request for the technological and economic growth of any nation. And the utilities must strive and ensure that the customers are getting the reliable power supply. In this thesis the reliability of jigjiga distribution system was done on both 15 and 33kv system assesses the performance of existing system. More over predictive reliability analysis for the future system considering optimal network reconfiguration to be annualized. For predictive reliability improvement, Matlab modeled fuzzy logic based optimization techniques has been applied, which delivers optimal number and placement of switches on the selected feeder of the distribution system. Among 8 feeders of the distribution system under cause study, 33kv feeder2 is selected as a test feeder as it has higest frequency and duration of interruption and also its cost of interruption is very high compared to other feeders .As the result of network reconfiguration in strategic places reduce the outage time in case of abnormal conditions and improve the reliability of the network and the optimal placement of switches and optimal tie switch of near by healthy feeder The SAIDI and SAIFI are reduced from 368.5failure/cust,yr to 166.37 failure/cust.yr and 621hr/cust.yr to 269.94hr/cust.yr .from the test results, the proposed technique present better interruption cost in comparation with that of existing system. 1,609,907.89ETB to 840,145.5 ETB reduction along with improvement in reliability indices. This method is implemented and tested on Togowuchale feeder of jigjiga town distribution network, Results of both initial and optimal configuration of expected energy not supplied (EENS) was improved from 1343MWH./yr to 701 MWh./yr this alternative which gives low system average interruption index (SAIDI), system average interruption frequency index(SAIFI) and customer average interruption duration index(CAIDI) was being assessed and considered. The reliability analysis indicated that of jigjiga distribution feeder was further improved reconfigured of the radial network in strategic place by using fuzzy logic artificial intelligence method. Keywords: Distribution Reliabilty, Network Reconfiguration, fuzzy logic optimization, distribution system reliability improvement