Massive MIMO systems offer a high spatial resolution that can drastically increase the spectral and/or energy efficiency by employing a large number of antennas at the base station(BS).In a distributed massive MIMO system,the capacity of fiber backhaul that links base station and remote radio heads is usually limited,which becomes a bottleneck for realizing the potential performance gain of both downlink and uplink.To solve this problem,we propose a joint antenna selection and user scheduling which is able to achieve a large portion of the potential gain provided by the massive MIMO array with only limited backhaul capacity.Three sub-optimal iterative algorithms with the objective of sumrate maximization are proposed for the joint optimization of antenna selection and user scheduling,either based on greedy fashion or Frobenius-norm criteria.Convergence and complexity analysis are presented for the algorithms.The provided Monte Carlo simulations show that,one of our algorithms achieves a good tradeoff between complexity and performance and thus is especially fit for massive MIMO systems. Massive MIMO systems offer a high spatial resolution that can drastically increase the spectral and / or energy efficiency by employing a large number of antennas at the base station (BS) .In a distributed massive MIMO system, the capacity of fiber backhaul that links base station and remote radio heads is usually limited, which becomes a bottleneck for realizing the potential performance gain of both downlink and uplink. To solve this problem, we propose a joint antenna selection and user scheduling which is able to achieve a large portion of the potential gain provided by the massive MIMO array with only limited backhaul capacity. Three sub-optimal iterative algorithms with the objective of sumrate maximization are proposed for the joint optimization of antenna selection and user scheduling, either based on greedy fashion or Frobenius-norm criteria. Legal and complexity analysis are presented for the algorithms. provided Monte Carlo simulations show that, one of our algorithms achieves a good tradeoff between complexity and performance and thus is especially fit for massive MIMO systems.