Anthropogenic interventions and actions upon the marine habitat pose threats to a range of species of economic and conservation concern. The dynamic nature of marine ecosystems offers a difficult challenge to incorporate spatial and temporal distributions of different species, and the interactions among species and human activities into a formal management framework. Each country has its own priorities when it comes to management of the marine resources (e.g. conservation, food security, sustainable fisheries, and optimization of revenue). Therefore, a key hurdle is to create tools adequate for use within an Ecosystem Approach to Management (EAM) and Ecosystem-based Management (EBM) framework, that meet local and regional needs. Models can provide insights regarding ecosystems dynamics and generate tools for management applications, including the estimation of optimal conditions and frameworks, assessing current conditions relative to baselines, exploring the effects of potential management decisions and delimiting areas where monitoring efforts of species of concern or “choke species” should be concentrated. The present work focuses on all these elements with the aim to provide modeling and visualization capacity to management decision making. My dissertation had two main objectives, divided in two case studies in distinct geographic and data availability settings. The first was to develop spatial models to promote the adequate monitoring of species of conservation concern (SOC) within a data-limited setting in two multiuse marine protected areas (MPAs) in the Amazon Delta, Northern Brazil by: 1) collecting SOC available data; 2) developing GIS-based suitability models; and 3) generating baseline knowledge for future management strategies of SOCs. The second objective was to study alewife (Alosa pseudoharengus) role within a data-rich setting in the Northeast U.S. large marine ecosystem (NEUS LME) and the Gulf of Maine (GOM) marine ecosystem by: 1) developing food-web based ecosystem models; 2) assessing the impacts of anadromous forage fish restoration; 3) testing riverine and marine-based management strategies to promote their recovery. For the data-limited studies, I found that extreme data scarcity impeded our ability to develop a model for the Environmental Protection Area (EPA) of Algodoal-Maiandeua, Northern Brazil. However, it allowed us to show preliminary data of sea turtles’ observations and fixed fishing gears in the EPA, giving basis to the future develop of spatially explicit models. While for the second multiuse MPA, the Soure Marine Extractive Reserve (MER), we were able to successfully develop a spatial explicit suitability model focused on monitoring priority areas for SOCs. Our results show that 30% of the MER is under medium, high and extremely high monitoring priority, allowing more effective development of monitoring design for SOCs. For the data-rich setting, I found that the full restoration of alewife in the NEUS LME could lead to a 50% potential biomass increase for small pelagics, 26% for fisheries target species, and approximately 69% for SOCs. This provided a more stable picture for the middle trophic level forage species and lead to major potential biomass changes for SOCs. I also found that fishing effort reduction alone did little to promote alewife recovery in the GOM marine ecosystem. However, when river to ocean connectivity was added to fishing effort reductions. The alosine (alewife, blueback herring, and American shad) group showed a major response. As a whole, my dissertation captures a range of management approaches from data-limited to data-rich systems, using modeling approaches to optimize decision making.