Advances in remote sensing is critical in assessing and monitoring ecosystem services in the context of agricultural productivity as affected by the impact of climate change that favours outbreak of new plant diseases. Remote sensing enables the quantification of threats to ecosystem support services in plant breeding. These include field-based phenotyping of maize diseases, constituting the identification, classification, quantification, and prediction. However, such support through sensor-based approaches are faced with multiple challenges when phenotyping diseases that appear simultaneously as co-infections on the same plant. Empirical evidence of sensor-based phenotyping and unmixing co-infections is non-existent. Furthermore, symptoms that cannot be categorized as co-infections also affect sensor-based phenotyping, e.g., symptoms similarities, different symptoms by same pathogen, and asymptomatic diseases. We review and discuss some of the challenges associated with sensor-based phenotyping of maize diseases under co-infections. We further suggest some possible practical application of unmixing co-infections using machine learning algorithms and suggest future efforts that may be possible in maize disease phenotyping using sensor-based approaches. The increasing capabilities of hyperspectral algorithms offer prospects for unmixing co-infections using sensor-based approaches. The thermal portion of the electromagnetic spectrum remains an active research area in phenotyping co-infections.