Abstract Pomegranate molasses is made from concentrated pomegranate juice with nothing added. Due to its nutritional value, limitation in production, and high production cost, this product may be adulterated by date syrup. This study was done to differentiate various types of pomegranate molasses and investigate the possibility of nonauthenticity detection in pomegranate molasses samples using the hyperspectral imaging (HSI) technique compared with physicochemical measurement analysis. The physicochemical properties (brix index, sucrose, acidity, ash content, pH, and formalin index) of 24 samples were measured as the reference analysis method, and it was found that the formalin index was a good factor for pomegranate molasses authenticity evaluation. Additionally, an HSI system (400–1000 nm) was used as a nondestructive and rapid screening method to capture spectral data of the samples. The evolutionary wavelength selection algorithm was applied to select effective wavelengths in sample clustering based on the obtained Davies–Bouldin index. Next, principal component analysis was used to visually interpret the spectral data of the sample when using the selected wavelengths and the whole spectra of the samples. Finally, an automatic clustering algorithm by the artificial bee colony as an unsupervised method was developed for the clustering of the authentic and nonauthentic samples. The method did not need descriptively labeled samples and obtained agreed satisfactorily with the degree of nonauthenticity in the samples. This study showed that the developed HSI technique coupled with an automatic clustering algorithm could detect date syrup nonauthenticity in pomegranate samples from the level of 5% adulteration.