Abstract: With the development of radiation shielding disciplines and the increasing requirements of industry, more detailed and accurate radiation shielding calculations are required. Due to the development of computer, Monte Carlo method can be applied in radiation transport simulations in various application scenarios to obtain detailed results under acceptable time consumption. Gamma photon transport in large space can also be simulated by Monte Carlo method to cope with needs of shielding deign in large space radiation environments. However, deep penetration problem arises in large-space photon transport simulations because the optical path between the source and the detector is 10 times larger than the mean free path, and the photon flux can decrease by more than 8 orders of magnitude, which means the efficiency of Monte Carlo simulation is low. Besides, the small size of target region makes it harder for Monte Carlo method to obtain results in target region. Although point detector and ring detector could cope with such small target problems, the deep-penetration problem can greatly influence the convergence of point detector and ring detector. In order to address the problems in the Monte Carlo simulation of large-space photon transport, a novel consistent adjoint driven importance sampling(CADIS)-ring detector (RD) hybrid method was proposed in this work. The ring detector estimator was employed as tally method to obtain photon flux at target position while the CADIS method was applied to automatically generate source-biasing parameters and weight-window parameters. The photon particles in Monte Carlo simulation can be biased sampled near the target position when using CADIS method, which is able to increase tally efficiency in target position. Generally, the CADIS method was applied with tracklength estimator and collision estimator, but it can be also used with the ring detector because there are no theoretical impediments in combination of the two methods. So, the CADIS method is also able to improve the convergence of the ring detector as expectation. To prove this, the CADIS-RD method was applied in a large-space photon transport problem and compared with the ring detector method. To meet the requirement of using ring detector estimator, the problem was modeled in cylindrical symmetry. In this cylindrical symmetry problem, the source is on the axis and the farthest target position is 3 000 m from the axis. As the results show, the CADIS-RD method performs better in variance reduction than the ring detector estimator in the same photon-transport situation. As for calculation of the photon flux at 3 000 m from the axis, compared with the ring detector method, the CADIS-RD method is able to reduce the relative standard deviation from 11.33% to 2.49%, which means a more precise result. Besides, the CADIS-RD method can improve the FOM about 2 times, which means higher efficiency. Moreover, according to the trend of relative standard deviation with particle number, the CADIS-RD method has better convergence of relative standard deviation than ring detector method. Thus, the novel approach in this work is able to handle the deep penetration problem and small target problem in large-space photon transport efficiently.