This study documented reliability of a picture identification instrument and determined its utility with young children in selected preschool programs. Three hundred eleven children, ages three-five, were selected from formal early childhood education programs in three states. A comprehensive health knowledge instrument, consisting of 30 items, was administered to all children, and a retest was administered two weeks later to a 20% randomly selected subsample. Data analysis included reliability assessments and systematic modification of the instrument. Pearson's Product Moment Correlation and Kuder Richardson 21 analyses indicated the instrument is reliable when administered individually to children ages three-five. The authors recommended this instrument, or a similar picture identification instrument, be used to assess comprehensive health knowledge.