<p>Feature selection for additive hazards models has gained increasing popularity in survival analysis in the presence of high-dimensional covariates. Despite the fast growing literature, most existing methods mainly focus on the aspect of statistical power and can not guarantee the reproducibility of the selection results. In this paper, we propose a new feature selection procedure to facilitate reproducible learning in high-dimensional additive hazards models by effectively controlling the false discovery rate (FDR) while maintaining high statistical power. We theoretically show that it can effectively control FDR in finite samples with an arbitrarily large number of covariates and the power of the proposed procedure approaches one asymptotically as the sample size increases infinitely. The effectiveness of the suggested method is demonstrated through simulation studies and two real-data analyses.</p>