Spatio-temporal variability of crop production strongly depends on soil heterogeneity, meteorological conditions, and their interaction. Canopy reflectance can be used to describe crop status and yield spatial variability. The objectives of this work were to understand the spatio-temporal variability of maize (Zea mays L.) yield using ground-based reflectance acquisitions in a salinity- and water-stress-affected 21-ha field beside the Venice Lagoon, Italy. Intra- and inter-annual reflectance variations were analyzed, across the entire field and at each map-cell over time, to understand how the different soil-related stress types (i.e., salinity and water) arise under different meteorological conditions. The results show that normalized difference vegetation index (NDVI), acquired during the maize flowering and kernel maturation stages (over the three growing seasons of 2010, 2011, and 2012), effectively describes yield spatio-temporal variability. In particular, stressed areas exhibited the smallest changes in NDVI over a single growing season. Soil salinity and water stress are responsible for ca. 44% of the intra-annual NDVI change. When multi-year NDVI data are compared, areas affected by soil salinity show the smallest temporal variability. Nevertheless, areas that are slightly saline and constantly affected by water stress could not be distinguished from highly saline areas. Multi-year reflectance data can be a useful tool to characterize areas where soil salinity is the main factor limiting crop production. In areas where several plant stresses occur simultaneously every year, the proposed approach could be used to guide precision irrigation to make adjustments for within-field leaching requirement and/or irrigation needs.