From September 1999 to March 2011, sentinel surveillance of measles was conducted by a self-selected sample of private physicians in Germany. From 2001, when mandatory surveillance for measles was established, two surveillance systems worked in parallel. The aim of this article is to summarize the strengths and limitations of sentinel versus mandatory surveillance. Active monthly reporting included case-based questionnaires on patients with (suspected) measles or zeroreporting. For confirmation of measles, the diagnostic patient specimens were sent to regional laboratories for serological tests or to the National Reference Laboratory (NRC). In the NRC in addition to serological tests measles-virus (MV) detection by PCR in urine, throat swabs, and oral fluid (since 2003) as well as MVgenotyping was offered. From January 2000 to December 2010, 934 out of 1,488 participating sentinel-practices did not see any measles case, while 554 reported 3,573 suspected cases. Measles was confirmed by laboratory testing in 801 cases, excluded in 473 cases, and the diagnosis remained uncertain in 215 cases. Of 3,100 analyzed cases, 2,712 (87 %) were unvaccinated, 217 (7 %) and 32 (1 %) were vaccinated with one or two doses, respectively, and for 139 (4 %) cases the vaccination status was unknown. The main reason for not being vaccinated against measles was refusal (n = 1,383). The confirmation rate was lower in the vaccinated than in the unvaccinated patients (19 % vs. 63 %). Since 2006, sentinel-cases have differed from notified cases by region and age. The proportion of sentinel cases from all NRC-investigated cases decreased from more than 50 % (2002) to less than 5 % (since 2007). Sentinel surveillance allowed for the detection of trends, delivered additional information for measles prevention, and played a major role in measles diagnostics. Since mandatory surveillance was established and sentinel surveillance no longer reflected the epidemiologic situation, the latter was abandoned in April 2011.