In this study, we employed a dual strategy for parathion organophosphate pesticide (parathion) detection; first, we used a localized surface plasmon resonance (LSPR)-based colorimetric sensor featuring thiol-capped Au NPs, namely cysteine (Cys)@Au NPs, 11-mercaptoundecanoic acid (11-MUA)@Au NPs, and glutathione (GSH)@Au NPs, via acetylcholinesterase (ACHE) and acetylthiocholine (ATCH) enzyme-mediated hydrolysis reactions; second, we developed a color analysis toxicity-sensing app (Toxin APP). Positively charged thiocholine (TCH) molecules, which were continuously generated via hydrolysis, subsequently conjugated with thiol-capped Au NPs, causing Au NP aggregation through electrostatic attractions. The degree of aggregation of the thiol-capped Au NPs was influenced by parathion concentrations in the range 0 to 108 ppt, because parathion acted as an ACHE inhibitor by controlling the amount of TCH generated. Based on the values of LSPR absorbance ratio, the limits of detection (LODs) of three types thiol-capped Au NPs were determined to be 100 ppt using ultraviolet-visible spectroscopy measurements. However, the aggregation efficiency of GSH@Au NPs was lower than that of the others regarding gradual changes in their color and LSPR absorbance band. Furthermore, we designed Toxin APP for color analysis which consists of three modules: processing, database collection, and communication. Toxin APP could on-site and precisely detect the color changes of GSH@Au NPs at parathion concentrations in the ranges of 100 ppt to 1, 10, and 100 ppm and could distinguish between OP and non-OP pesticides (e.g., fipronil) in tap water samples with high sensitivity and selectivity. Moreover, the concentration of residual parathion in real samples (tomato and strawberry) was quantified based on the color changes of GSH@Au NPs detected using Toxin APP. Therefore, the combination of an LSPR-based colorimetric assay and Toxin APP can be a reliable method for the facile and rapid detection of parathion in food and water samples.