This research was designed to determine and compare the technical efficiency and input levels used in rice production under farmer managed irrigation systems (FMIS) and rain fed systems (RFS) in Kogi State. It also compared the effects of socioeconomic characteristics on the technical efficiency of farmers in the FMIS and RFS. Four null hypotheses were tested. The study was conducted in commercial rice producing areas of Kogi State. It adopted a multi stage purposive sampling technique. Agricultural Zones where rice is produced in commercial quantities were purposively stratified into three (3) based on a preliminary survey. From these three zones, one local government area (LGA) each was selected based on the availability of commercial rice farms in the area. Out of these LGAs (Ibaji, Bas a and Kogi LGAs), forty (40) rice farmers each were randomly sampled giving a total sample size of one hundred and twenty (120) rice farmers. Primary data were obtained by interviews via a set of structured questionnaires. Data were analyzed using descriptive statistics, Levene’s test, Welch and Brown-Forsythe robust tests for equality of means, Chow-break point test and maximum likelihood estimation (MLE) of stochastic frontier and inefficiency models. The mean age of farmers in the study area was 42 years. The farmers in the study area spent a mean of 8 years on formal education. Seventy two percent (72%) of the farmers were males while twenty eight percent (28%) were females. Women were not participating remarkably well especially in ownership of rice farms in the study area. The mean value of rice farming experience in the study area was 16years. Results showed that the FMIS had a higher intensity of inputs usage than the RFS. In the input comparison between FMIS and RFS, statistically significant positive mean differentials were recorded for land, fertilizer quantities applied, family and hired labour, quantities of pesticides used on the farm and value of water used on the farm per farming season. The estimated elasticities of mean output with respect to land, fertilizer, family labour, seeds, and water were statistically significant at less than 1 percent and 5 percent in the FMIS. Their respective elasticities were 0.33, 0.010, 0.075, 0.151 and 0.165. It was indicated that land size (farm size) and quantities of fertilizer applied by the farmers, were the statistically significant determinants of technical efficiency in the RFS. The elasticities of rice output with respect to the inputs, land and chemical fertilizer utilized were 0.276 and 0.024 respectively. This result is unlike the FMIS where five variables had statistically significant elasticities. The mean technical efficiency of the FMIS was 73 percent. It was lower than that of the rainfed system which had 90 percent. Significant difference existed in the technical efficiencies of the two groups. The returns to scale estimated, 0.813, and 0.476 for both FMIS and RFS respectively indicated that farms in the study area were characterized by decreasing returns scale. Farming experience, years of formal education and frequency of extension contacts exerted statistically significant effects on the technical efficiencies of the FMIS. Meanwhile four out of the six socio-economic variables, education, extension contact and age of farmers had statistically significant t-ratios or influences on the levels of rice output recorded by the RFS farmers. They were all significant at less than 1 percent alpha level. Significant differences existed in most of the socioeconomic variables of the two group of rice farmers studied in Kogi State. Five major recommendations were made which included the need for capacity building among farmers and extension agents, public investment in irrigation projects, public-private partnership aimed at encouraging resource conservation and inputs supply (including microcredit) to rice growing communities among others.