Coastal environments benefit from the movement and exchange of nutrients facilitated by water flows. While this process is important for mangroves, seagrass patches, and coral reefs found in tropical coastal environments, water flows can also play a major role in the detection and tracking of pollutants, conservation efforts, and applications of aquatic herbicides for managing submerged plants. Monitoring of water flows is difficult due to their complex and temporally dynamic movement. The domain of high frequency or continuous tracking of dynamic features such as water flows has previously been limited to in situ monitoring installations, which are often restricted to small areas or remote sensing platforms such as aircrafts, which are generally prohibitively costly. However, unmanned aerial vehicles (UAV) are suitable for flexible deployment and can provide monitoring capabilities for continuous data collection. Here, we demonstrate the application of a UAV-based approach for tracking coastal water flows via fluorescent dye (Rhodamine WT) released in two shallow-water locations in a coastal tropical environment with mangroves, seagrass patches, and coral reefs along the shores of the Red Sea. UAV-based tracking of the dye plumes occurred over the duration of an ebbing time. Within the first 80 min of dye release, red-green-blue UAV photos were collected at 10-second intervals from two UAVs, each hovering at 400 m over the dye release sites. Water samples for assessment of dye concentration were also collected within 80 min of dye release in 30 different locations and covered concentrations ranging from 0.65 - 154.37 ppb. As the dye plumes dispersed and hence covered larger areas, nine UAV flight surveys were subsequently used to produce orthomosaics for larger area monitoring of the dye plumes. An object-based image analysis approach was employed to map the extent of the dye plumes from both hovering UAV photos and the orthomosaics that were geometrically corrected based on GPS-surveyed ground control points and radiometrically corrected based on black, grey, and white reflectance panels. Accuracies of 91 – 98% were achieved for mapping dye plume extent when assessed against manual delineations of the dye plume perimeters. UAV data collected coincidently with the water samples were used to predict dye concentrations throughout the duration of the ebbing tide based on regression analysis of band indices. The multiplication and the red:green and red:blue ratios provided a best-fit regression between the 30 field observations of dye concentration and the 30 coincident UAV photos collected while hovering with a coefficient of determination of 0.96 and a root mean square error of 7.78 ppb. The best-fit equation was applied to both the hovering UAV photos and the orthomosaics of the nine UAV flight surveys to detect dye dispersion and the movement of the dye plume. At the end of the ebbing tide, the two dye plumes covered an area of 9,998 m2 and 18,372 m2 and had moved 481 m and 593 m, respectively. Our results demonstrate how a UAV-based monitoring approach can be applied to address the lack of understanding of coastal water flows, which may facilitate more effective coastal zone management and conservation.