One of the major applications of the CRISPR-Cas9 system is the visualization of DNA by using nuclease-deactivated Cas9 (dCas9), which, following complexation with a single-guide RNA (sgRNA), specifically binds to target genomic sequences without inducing DNA breaks. In this approach, either dCas9 or the sgRNA is labeled with fluorescent proteins or dyes, or they are engineered to recruit such molecules. A key advantage of CRISPR-based imaging is that genomic elements in living cells can be tracked, because of the ability to express all system components in vivo. Although CRISPR-based imaging has been successfully used to label repetitive sequences in living cells, the visualization of nonrepetitive loci remains a challenging issue. The primary obstacles are a low signal-to-noise ratio and the potential for nonspecific DNA binding by the dCas9-sgRNA complex, which can generate fluorescent puncta at off-target sites. Efficient intracellular delivery of system components and their sustained expression over time are also a major concern. Consequently, CRISPR-based imaging remains a highly time- and labor-intensive process that requires ongoing optimization. Here, we summarize recent advances in labeling nonrepetitive genomic loci, outline key challenges associated with CRISPR-based imaging, and present insights derived from our own experimental findings and research experience.