This paper investigates the non-orthogonal multiple access (NOMA) based multicasting in wireless caching networks for the performance enhancement of spectral efficiency. Particularly, at the base stations (BSs), upon receiving the requests from the attached user equipments (UEs), a superposed message of the most requestedm files is broadcasted by applying NOMA to simultaneously satisfy the demands of multiple UEs, with the power allocation of the files determined by the respective number of requests. Further, at the UEs, two different decoding strategies, i.e., the conventional SIC without cache-aided interference cancellation and the SIC with cache-aided interference cancellation, are considered to facilitate the reliable data reception. To analyze the performance of the proposed NOMA based multicasting scheme, a novel framework is developed based on stochastic geometry. We first derive the probability that an ordered sequence Y ofm files is broadcasted by the typical BS Y with the technique of NOMA. Then, we characterize the coverage probability of i-th file in Y. Further, we evaluate the average coverage probability of the wireless caching network with the studied NOMA based multicasting. It is shown by simulations that the derived analytical results are validated and the technique of wireless caching can further improve the performance of NOMA.