The recycling of plastics greatly contributes to the preservation of the natural resource oil and the reduction of the flow of plastic waste to our overflowing landfills. Plastic recycling is highly attractive not only environmentally, but also from an economic point of view. As a result, a rapidly growing interest in plastic recycling technologies exists. An important requirement for high-quality recycling products is the high purity of the recycled material which makes an automated, reliable identification and sorting process of different plastics indispensable. We propose to label the different plastic resins using n fluorescent markers in a way such that unique spectral signatures are generated and thereby avoid the known drawbacks of existing plastic sorting methods. These markers must neither change the properties nor the visual appearance of the plastic parts and are thus to be incorporated with concentrations of only a few ppm on a molecular level during manufacture. The aim of this study is to evaluate the classification performance as a function of the signal-to-noise ratio (SNR) and the computational load of 5 algorithms known to be successfully applied to other applications plus a novel approach when using 4 specific fluorescence markers currently at our disposal. With a simple binary coding scheme 24−1 = 15 different fluorescence emission spectra can be generated which represent 15 different sorts of fluorescently labeled plastics. These investigations shall help optimize the overall measurement system with regard to e.g. the fluorescence excitation source and optical sensors, the spectral shape of the fluorescence emissions and hence the design of more fluorescence markers, etc.