Diet studies generally rely on conventional methods with inherent technical limitations which rule out analyses on carnivores mixed diets and inferences about individual foraging variation. Here, using DNA-based methods, we investigated the effects of feeding degrees and seasons on the diet of rural free-ranging domestic cats. First, we evaluated the detection of food item DNA by amplifying a cytochrome b minibarcode from collected faeces of a housebound cat fed a known diet and by sequencing amplicons using two next-generation sequencing (NGS) benchtop platforms (Ion Torrent PGM and Illumina MiSeq) to validate our approach. Prey DNA was better detected than DNA associated with anthropogenic foods and the sensitivity of detection of food items differed according to NGS platforms. Then, we used DNA metabarcoding on 359 a priori cat faeces collected during three sessions and combined results from the two NGS platforms. Simultaneously, the producer of faeces was identified by comparing individual microsatellite genetic profiles from faeces and fur samples of outdoor cats fed ad libitum (n = 86) and farm cats irregularly fed (n = 68). Of collected faeces, 90.8 % were confirmed to be that of domestic cats. At least one food item was determined in 81.9 % of the investigated faeces, and 14.4 % of the producer of faeces were assigned with genotyping. A total of 24 different prey taxa were identified, with 93 % until the species level. Rodents were most frequently eaten by cats (90.1 %), followed by birds (9.5 %). Microtine voles (Arvicola terrestris and Microtus arvalis) occurred preferentially in cat diet. During the cool season, the consumption of preys by outdoor cats was significantly lower than the one of farm cats, while it did not differ during the mild season. Feeding degree and weather influenced cat predation. This DNA-based approach appears as a very promising tool to assess the diet of rare and elusive carnivores or the exploitation of anthropogenic foods by urban wildlife species.