Masiza, WongaNkuna, Basani LammyRatshiedana, Phathutshedzo EugeneMadasa, AkhonaNduku, LwandileShwatja, TumeloChirima, Johannes GeorgeNyamugama, AdolphAbutaleb, KhaledKhoboko, Pitso WalterHamandawana, Hamisai2026-03-192026-03-192026-03Masiza, W., Nkuna, B.L., Ratshiedana, P.E. et al. 2026, 'Predicting smallholder maize yield using sentinel-2-derived phenological metrics', Smart Agricultural Technology, vol. 13, art. 101870, pp. 1-12, doi : 10.1016/j.atech.2026.101870.2772-3755 (online)10.1016/j.atech.2026.101870http://hdl.handle.net/2263/109076DATA AND CODE AVAILABILITY : The Python code used for data processing, modelling, and feature-importance analysis in this study is publicly available at: https://github.com/masizawonga63-eng/sentinel2-maize-yield-phenology. Due to farmer confidentiality agreements and ethical restrictions, the raw field-level yield data and associated farm identifiers cannot be publicly shared. Derived phenological metrics and aggregated results supporting the findings of this study are available from the corresponding author upon reasonable request.Please read abstract in the article. HIGHLIGHTS • Phenological metrics derived from multiple spectral indices are used to predict maize yields. • Regularized linear models were trained with limited data to predict maize yields. • Pre-peak and cumulative integrals of red-edge indices best predicted maize yield. • Parsimonious models trained with key features showed no measurable loss of accuracy.en© 2026 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).SmallholderPhenologyRemote sensingCrop yieldMaizeArticle