This paper describes a colorimetric sensor array for the discrimination of volatile amines. Analyte discrimination is achieved by combining two functional elements: (1) a "chemical class-selective" single chromogenic sensing dye with selectivity for amines in general, encapsulated into (2) polymer nanoparticles with different polarities. The resulting array has the ability to distinguish one closely related amine from another, relying on a polarity-based approach. In order to achieve reproducible, cost efficient, and flexible sensor array fabrication with the potential for mass production, inkjet-printing technology combined with standard copy paper as a sensor substrate is applied. Printing of 6 types of inks, which are prepared by mixing two dye encapsulating nanoparticles of different polarity in different mixture ratios, results in a colorimetric sensor array with a polarity gradient. Seven primary amines with increasing alkyl chain lengths have been selected to demonstrate the performance of the sensor array. The RGB color differences (ΔR, ΔG, ΔB) of the sensor array spots before and after gas exposure were analyzed by principal component analysis (PCA) and agglomerative hierarchical clustering (AHC) analysis. Under the selected measurement conditions, results of PCA and AHC analysis indicated high discrimination ability with high reproducibility of the sensor array down to amine concentrations of 50 ppm. The discrimination ability was maintained at relative humidities between 10% and 80%. Furthermore, the sensor array showed no significant response to common volatile organic compounds, confirming the high selectivity toward amines. This is, to the best of our knowledge, the first report of a colorimetric sensor array with selectivity for a specific chemical class of analytes and the ability to discriminate compounds of the same class, which is obtained by simply mixing two types of single dye-encapsulating polymer nanoparticles.