Citrus fruit is highly susceptible to postharvest decay caused by fungal and oomycete pathogens, emphasising the need for non-destructive early detection tools operating directly on packaged fruit and headspace. In this study, a smartphone-read fluorescent sensor array was evaluated as a portable cross-reactive platform for detecting citrus postharvest pathogens and capturing infection-associated chemical fingerprints through chemometric pattern analysis. An array of 17 cross-reactive fluorescent probes was spotted on polyamide filters and imaged under 365 nm illumination. Spore suspensions of Penicillium digitatum, P. italicum, Alternaria alternata, Phytophthora citrophthora and Ph. nicotianae (10²–10⁶ propagules mL⁻¹) were analysed. Partial least squares (PLS) calibration for both Penicillium species showed linear trends, with R² = 0.9999 and 0.9987, supporting quantitative estimation of propagules in aqueous suspensions. PLS-discriminant analysis of array responses yielded three well-separated clusters corresponding to Alternaria, Penicillium and Phytophthora. In vivo assays on oranges inoculated with P. digitatum showed that the array captured changes in headspace composition, distinguishing infection from a physiological wound response. While healthy wounded oranges followed a consistent trajectory, inoculated oranges showed a clear divergence as early as 24–48 h post-inoculation, prior to visible symptom onset. A dedicated model for infected oranges yielded a linear relationship between storage time and predicted value over 0–4 days. These findings indicate that smartphone-read cross-reactive fluorescent arrays can provide early, non-destructive screening of citrus postharvest decay and support risk-oriented prioritization of lots for subsequent targeted confirmatory analyses, including pathogen-associated metabolites of food safety relevance.