Chinese hamster ovary (CHO) cells are the most important platform for producing biotherapeutics. Random integration of a transgene into epigenetically instable regions of the genome results in silencing of the gene of interest and loss of productivity during upstream processing. Therefore, cost- and time-intensive long-term stability studies must be performed. Site-specific integration into safe harbors is a strategy to overcome these limitations of conventional cell line design. Recent publications predict safe harbors in CHO cells based on omics data sets or by learning from random integrations, but those predictions remain theory. In this study, we established a CRISPR/Cas9-mediated site-specific integration strategy based on ChIP-seq data to improve stability of recombinant CHO cells. Therefore, a ChIP experiment from the exponential and stationary growth phase of a fed-batch cultivation of CHO-K1 cells yielded 709 potentially stable integration sites. The reporter gene eGFP was integrated into three regions harboring specific modifications by CRISPR/Cas9. Targeted Cas9 nanopore sequencing showed site-specific integration in all 3 cell pools with a specificity between 23 and 73%. Subsequently, the cells with the three different integration sites were compared with the randomly integrated donor vector in terms of transcript level, productivity, gene copy numbers and stability. All site-specific integrations showed an increase in productivity and transcript levels of up to 7.4-fold. In a long-term cultivation over 70 generations, two of the site-specific integrations showed a stable productivity (>70%) independent of selection pressure.