Man-made reservoirs are important inland ecosystems that provide food and livelihoods in many countries. Due to the dispersed nature of inland fisheries, most individual systems are rarely adequately assessed or monitored, therefore reliable data on the target stocks are largely unavailable to implement management strategies. This thesis focuses on Tono, Bontanga and Golinga reservoirs in northern Ghana which contribute significantly to food nutrition and community livelihoods. The thesis presents studies that demonstrate how differences in reservoir use patterns, reservoir morphometry and physicochemical characteristics influence ecosystem structures and fisheries resource productivity. The thesis includes (i) an assessment of empirical models for reservoir harvest estimations, (ii) a study relating morphometric characteristics of the reservoirs to fish production, (iii) assessments of the exploitation levels and stock status of the reservoirs’ target species and (iv) a holistic description of the reservoirs’ biological interactions through a food web modelling approach (Ecopath with Ecosim). To improve the estimation of current harvest potential of West African reservoirs, the relationship between total annual fish catch and reservoir surface area was modelled, which compared to a previous model, indicated that catches from reservoirs in the region have more than doubled over the last two decades. While the analysis indicated that fisheries productivity is inversely correlated with both mean depth and surface area, no significant correlation was found with reservoir age. The exploited resources in the small-scale fisheries of the Tono, Bontanga, and Golinga reservoirs were assessed based on length frequency samples. Growth, mortality, exploitation status, stock size, and relative yield per recruit reference points were determined using bootstrapping fish stock assessment (BFSA), a novel framework that allows for the estimation of uncertainties around the life-history parameters and reference levels (e.g., L∞, K, and F0.1). A complementary assessment approach based on length-based indicators was used to calculate the species’ spawning potential ratios under the current exploitation regime. Tono, Bontanga and Golinga reservoirs provide a total fish catch of 10.1, 15.5 and 17.1 t km−2 yr−1, respectively. The reservoir fisheries are dominated by two cichlid species (Sarotherodon galilaeus and Oreochromis niloticus). The cichlid species Oreochromis niloticus, Sarotherondon galilaeus, and Coptodon zillii were found to be heavily exploited in all three reservoirs. The giraffe catfish, Auchenoglanis occidentalis was found only in Tono and Bontanga reservoirs. In Bontanga, the catfish stock is fully exploited. While in Tono, the giraffe catfish is underexploited, the current fishing mortality could be doubled to increase yield. The length-based indicators suggested all the species at Bontanga and O. niloticus and S. galilaeus populations at Golinga have spawning stock biomasses below 40% of the unfished biomass. This points to a situation of a possible ongoing recruitment overfishing of those species in the two reservoirs and suggests that a further increase in fishing effort is not advisable. To support the construction of reservoirs’ food web models, a study was conducted on the feeding characteristics of the giraffe catfish with the expectation that the population in Reservoir Tono, which has an extensive macrophyte coverage, feeds more on plant material and associated insects than their counterparts in the Reservoir Bontanga. The study showed that fish food items did not differ significantly between the two reservoirs. Insect larvae and algae dominated the stomach contents. Comparative analysis of the reservoirs showed interesting differences: the mean trophic level of the catch was lowest in the largest and deepest reservoir (Tono), likely due to higher trophic level species occupying less accessible deep ‘refuge’ habitats. In the medium-sized (Bontanga) and small shallow (Golinga) reservoirs, in contrast, a larger catch portion resembles high trophic level species. Lake Bontanga differs from the other reservoirs by having a lower human population impact, a significantly lower Total Primary Production to Total Respiration ratio, a higher Total Biomass to Total System Throughput ratio, a higher Finn Cycling Index, a higher Detritivory to Herbivory ratio as well as the highest gross efficiency of the catch, all indicative for a more developed ecosystem. The smallest shallow (Golinga) reservoir is more impacted by anthropogenic activities than the other two reservoirs as indicated by the high levels of dissolved organic carbon, total dissolved nitrogen bonded, nitrite-nitrogen and turbidity in the reservoir. While the smallest lake had the highest fish production (per unit area) under optimal conditions of water supply, it is most vulnerable when used for both irrigated agriculture and fisheries production. The findings of this thesis suggest that the use of man-made lakes and respective catchment areas should be assessed and managed carefully to prevent the loss of nutrition and livelihoods contributions. Finally, this thesis serves as a broad template for the development of sustainable ecosystem-based management measures not only for the three studied ecosystems but for other reservoirs exposed to human activities around the world.