Introduction. Surveillance for laboratory-confirmed influenza-associated deaths in children is used to monitor the severity of influenza at the population level and to inform influenza prevention and control policies. The goal of this study was to better estimate pediatric influenza mortality in New York state (NYS). Methods. Death certificate data were requested for all passively reported deaths and any pneumonia and influenza (P&I) coded pediatric deaths occurring between October 2004 and April 2010, excluding New York City (NYC) residents. A matching algorithm and capture-recapture analysis were used to estimate the total number of influenza-associated deaths among NYS children. Results. Thirty-four laboratory-confirmed influenza-associated pediatric deaths were reported and 67 death certificates had a P&I coded death; 16 deaths matched. No laboratory-confirmed influenza-associated death had a pneumonia code and no pneumonia coded deaths had laboratory evidence of influenza infection in their medical record. The capture-recapture analysis estimated between 38 and 126 influenza-associated pediatric deaths occurred in NYS during the study period. Conclusion. Passive surveillance for influenza-associated deaths continues to be the gold standard methodology for characterizing influenza mortality in children. Review of death certificates can complement but not replace passive reporting, by providing better estimates and detecting any missed laboratory-confirmed deaths.