Background. Influenza-like illness (ILI) is monitored in Guadeloupe and Martinique by a network of sentinel physicians. During the 2009 influenza A(H1N1) pandemic, the incidence of ILI estimated from these networks was higher in Martinique than in Guadeloupe, even though the demographic characteristics of the two islands are comparable. In October 2009, a telephone survey of the population on both islands was conducted to estimate the number of people who had been diagnosed with influenza during a medical consultation since May 2009. This study compares the results of sentinel surveillance, the telephone survey, and hospital surveillance conducted during the epidemic between the two islands. Methods. The Guadeloupean and Martinican sentinel networks consist of a sample of volunteer general practitioners, geographically distributed across each territory, who report weekly numbers of influenza-like illnesses. Time series of the estimated weekly number of cases were modeled using a Serfling-type seasonal component regression. The upper bound of the 95% one-sided confidence interval of the expected mean value constitutes the epidemic threshold for the week in question. The telephone survey was conducted from October 2 to 12, 2009, in Martinique and from October 13 to 21, 2009, in Guadeloupe, using a computer-assisted telephone interviewing system with 507 individuals in Martinique and 508 in Guadeloupe selected by the quota method and representative of individuals aged 15 and older. Results. The epidemic lasted 12 weeks on both islands, from August 3 to October 25 in Martinique and from August 17 to November 8 in Guadeloupe. Between August and September, the attack rate was estimated in Martinique at 5.52% (95% CI: 5.23–5.83) by the sentinel network, compared with 8.3% (95% CI: 6.0–11.0) by the telephone survey. In Guadeloupe, it was 2.13% (95% CI: 1.97–2.24) according to the sentinel network and 6.9% (95% CI: 4.8–9.5) according to the telephone survey. The number of confirmed hospitalized cases was equivalent on both islands. Conclusion. These results suggest a tendency toward underestimation by the sentinel network in Guadeloupe. However, this finding must be qualified by the results of hospital surveillance. The number of emergency department visits continues to point toward a higher incidence in Martinique, where the lower rate of testing hospitalized cases with influenza-like illness could partly explain the differences observed in hospitalization rates, severity, and case fatality rates during the epidemic between the two islands.