In San Francisco, men who have sex with men (MSM) with early syphilis are at high risk of reinfection. We described syphilis testing behavior among MSM after diagnosis, identified factors associated with not testing, and developed algorithms to identify nontesters. We used syphilis surveillance data from 2005 through 2008 to describe follow-up testing behavior among MSM with early syphilis and titers of non-treponemal serologic tests ≥1:16. We analyzed data from contact-tracing interviews to identify factors associated with not testing during the 1 to 6 months post-diagnosis. We developed and applied a multivariate model in a derivation set (2005-2006) and a validation set (2007-2008), respectively, calculating correct classification rates (CCR) to assess predictive ability and evaluating patient characteristics for potential interventions. Among 795 MSM, 260 (33%) did not have a follow-up syphilis test. Not testing was associated with being HIV-uninfected (risk ratio [RR]: 1.9, 95% confidence interval [CI]: 1.5-2.6), residing outside of San Francisco's gay-identified neighborhood (RR: 1.7, 95% CI: 1.0-2.9), and being diagnosed at the municipal sexually transmitted disease clinic (RR: 1.5, 95% CI: 1.2-2.0) (CCR derivation set, 71.6%; CCR validation set, 71.3%). An intervention focusing on MSM with those 3 characteristics would include 13% of syphilis cases among MSM and identify 26% of nontesters. Although MSM in San Francisco are at high risk for syphilis reinfection, one-third of MSM diagnosed with syphilis did not test during the 1 to 6 months post-diagnosis. Interventions to encourage follow-up testing among persons with syphilis might contribute to more effective syphilis prevention and control efforts.