Cyclooxygenase (COX)-derived prostaglandins (PGs) exhibit manifold functions in acute and chronic skin inflammation induced by a number of physical (ultraviolet (UV) light, wounding) and chemical (12-O-tetradecanoylphorbol 13-acetate (TPA), arachidonic acid) noxious stimuli. Depending on the challenge and the context, constitutively expressed COX-1 or the transiently induced COX-2 isoform are of relevance. Moreover, squamous cell carcinoma (SCC) of skin is a prominent example of epithelial neoplasia that consistently overexpresses COX-2 in the parenchyme and the mesenchyme of premalignant and malignant lesions, while COX-1 expression remains unaltered. Pharmacological, clinical, and experimental animal studies as well as a few epidemiological studies document the importance of PG signaling in non-melanoma skin cancer including SCC and basal cell carcinoma (BCC) in humans and mice. Increased levels of PGE₂ and PGF_2α in premalignant and/or malignant epithelial skin cancers are due to the constitutive upregulation of enzymes involved in PG biosynthesis, such as COX-2, and downregulation of the tumor suppressor gene 15-hydroxy-prostaglandin dehydrogenase (15-PGDH), which is involved in the inactivation of PG, thus counteracting the activities of COX. Most remarkably, genetic studies show that mice which are deficient in COX-2 or COX-1 are protected from the development of SCC when applying the multi-stage chemical carcinogenesis protocol. Conversely, the forced overexpression of COX-2 in the proliferative basal compartment of the stratified skin epidermis results in spontaneous hyperplasia and dysplasia in transgenic mice and furthermore a sensitization for cancer development by conferring an auto-promoted skin phenotype. In multi-stage carcinogenesis, it also becomes clear that aberrant COX-2 overexpression and activity are causally involved in tumor promotion and tumor progression rather than initiation. In contrast, using as inducer of carcinogenesis the complete carcinogen UV B light, depletion of COX-2 but not of COX-1 makes mouse skin resistant for SCC, indicating that here, only COX-2 is essential. Depending on the type of challenge, COX-2-dependent signaling contributes to the pre-invasive growth of the skin epidermis by a delayed onset of terminal differentiation, or stimulation of hyperproliferation and survival. With respect to BCC, the genetic ablation of COX-2 but also of COX-1 leads to a strongly reduced tumor burden in the skin of Patched (Ptch)1^+/− mice, which due to the deletion of a Ptch1 allele, spontaneously develop BCC resembling human familial basal cell nevus syndrome and sporadic BCC. Nonsteroidal anti-inflammatory drugs and the COX-2-selective inhibitors (COXibs) exhibit impressive efficacy inhibiting tumor burden in various mouse models of SCC and BCC. Most importantly, in humans the interruption of COX-2 signaling is an effective strategy to treat and chemo-prevent non-melanoma skin cancer in individuals who are at high risk for the disease. However, any potential beneficial effect of this medicine has to be balanced against the adverse effects that are known to be associated with these drugs in a subset of patients.