Background  To record and assess the clinical features of chikungunya fever (CHIKF), with a view to enable diagnosis based on clinical criteria rather than costly laboratory procedures in field conditions. Methods  As part of a cross-sectional serologic survey conducted in Mayotte after a massive chikungunya outbreak in 2006, we collected data on clinical features of chikungunya infection and assessed the performance and accuracy of clinical case definition criteria combining different symptoms. Results  Of 1154 participants included, 440 (38.1%) had chikungunya-specific IgM or IgG antibodies by enzyme-linked immunosorbent assay (ELISA). Of symptomatic participants, 318 (72.3%) had confirmed chikungunya, the dominant symptoms reported were incapacitating polyarthralgia (98.7%), myalgia (93.1%), backache (86%), fever of abrupt onset (85%) and headache (81.4%). There was a strong linear association between symptomatic infection and age (χ2 for trend = 9.85, P \textless 0.001). Only 52% of persons with presumptive chikungunya sought medical advice, principally at public primary health care facilities. The association of fever and polyarthralgia had a sensitivity of 84% (95% CI: 79–87) and a specificity of 89% (95% CI: 86–91). This association allowed to classify correctly 87% (95% CI: 85–89) of individuals with serologically confirmed chikungunya. Conclusions  Our results suggest that the pair fever and incapacitating polyarthralgia is an accurate and reliable tool for identifying presumptive CHIKF cases in the field. These criteria provide a useful evidence base to support operational syndromic surveillance in laboratory-confirmed chikungunya epidemic settings.