Indoor localization based on the fifth generation (5G) wireless communication technology has drawn much attention owing to its potential for supporting location based services. In this paper, we investigate fingerprint-based localization in 5G indoor environment and propose a 5G indoor intelligent localization scheme. The proposed scheme takes the advantage of the large bandwidth and subcarrier spacing of 5G signals to improve the localization accuracy. First, MUSIC algorithm is applied for angle of arrival (AoA) estimation based on the channel state information (CSI) matrix. Both the estimated AoAs and the corresponding amplitudes are collected as fingerprints. Second, deep neural network (DNN) is employed for location estimation. Simulation results show that the proposed scheme performs well in terms of localization accuracy and stability. The mean localization accuracy of the proposed scheme is better than 1.1 m even in the non-line-of-sight environment.