Artificial drainage is required to make wet soils productive for farming. However, drainage may have unintended environmental consequences, for example, through increased nutrient loss to surface waters or increased flood risk. It can also have implications for greenhouse gas emissions. Accurate data on soil drainage properties could help mitigate the impact of these consequences. Unfortunately, few countries maintain detailed inventories of artificially-drained areas because of the costs involved in compiling such data. This is further confounded by often inadequate knowledge of drain location and function at farm level. Increasingly, Earth Observation (EO) data is being used map drained areas and detect buried drains. The current study is the first harmonised effort to map the location and extent of artificially-drained soils in Ireland using a suite of EO data and geocomputational techniques. To map artificially-drained areas, support vector machine (SVM) and random forest (RF) machine learning image classifications were implemented using Landsat 8 multispectral imagery and topographical data. The RF classifier achieved overall accuracy of 91% in a binary segmentation of artifically-drained and poorly-drained classes. Compared with an existing soil drainage map, the RF model indicated that ~44% of soils in the study area could be classed as “drained”. As well as spatial differences, temporal changes in drainage status where detected within a 3 hectare field, where drains installed in 2014 had an effect on grass production. Using the RF model, the area of this field identified as “drained” increased from a low of 25% in 2011 to 68% in 2016. Landsat 8 vegetation indices were also successfully applied to monitoring the recovery of pasture following extreme saturation (flooding). In conjunction with this, additional EO techniques using unmanned aerial systems (UAS) were tested to map overland flow and detect buried drains. A performance assessment of UAS structure-from-motion (SfM) photogrammetry and aerial LiDAR was undertaken for modelling surface runoff (and associated nutrient loss). Overland flow models were created using the SIMWE model in GRASS GIS. Results indicated no statistical difference between models at 1, 2 & 5 m spatial resolution (p< 0.0001). Grass height was identified as an important source of error. Thermal imagery from a UAS was used to identify the locations of artifically drained areas. Using morning and afternoon images to map thermal extrema, significant differences in the rate of heating were identified between drained and undrained locations. Locations of tiled and piped drains were identified with 59 and 64% accuracy within the study area. Together these methods could enable better management of field drainage on farms, identifying drained areas, as well as the need for maintenance or replacement. They can also assess whether treatments have worked as expected or whether the underlying saturation problems continues. Through the methods developed and described herein, better characterisation of drainage status at field level may be achievable.