Organic matter preserves well under desiccation. Therefore, textiles may be found at archaeological sites in arid environments. The aim of the present study was to develop a rapid and cost-effective method to assign hair and wool in ancient textiles to animal species. We created a reference based on light microscopy of modern hair samples, relevant to the study region, including two sheep breeds, dromedary and Bactrian camels, goats and rabbits. The method combined several hair features for accurate species identification: medullation rate, cuticular attributes, fiber diameter, and the ratio of scale length to fiber diameter. Statistical models ascertained the distinction between sheep and non-sheep fiber to a 90% level of confidence. Distinguishing between goat and camel was challenging but, after sheep was withdrawn from the model, the probability of belonging to goat was evaluated to an 80% level of confidence for very coarse hairs. Features that distinguish rabbit hairs unequivocally were identified. A customized work-flowchart was instrumental in studying of 20 specimens of hair, wool or fleeces, mostly from sheep, and strings, from the Roman, Byzantine and Early Islamic periods. Overall, this multi-proxy method allows for confident identification of animal species in ancient fabrics from the arid southern Levant.