Background: Estimating dietary intake is important for both epidemiological and clinical studies, but often lacks accuracy. Objective: To investigate the accuracy and validity of energy intake estimated by an easy-to-use semiquantitative food record (EISQFR) compared to total energy expenditure (TEE) estimated by doubly labelled water technique (EEDLW). Design: TEE was measured in 29 nonobese subjects using the doubly labelled water method over a period of 14 days. Within this period, subjects reported their food consumption by a newly developed semiquantitative food record for 4 consecutive days. Energy intake was calculated using the German Food Code and Nutrition Data Base BLS II.3. Results: A good correlation was observed between EISQFR and EEDLW (r = 0.685, P<0.001). The mean difference between EISQFR and EEDLW was -1.7±2.6MJ/day (-14±21% P = 0.002). An underestimation of EISQFR <10% was observed in nine subjects (31%), of 10-20% in six subjects (21%), and of >20% in nine subjects (31%). In five subjects (17%), an overestimation of EISQFR was observed. Conclusions: The easy-to-use semiquantitative food record provided good estimates of EI in free-living and nonobese adults without prior detailed verbal instructions. The presented food record has limitations regarding accuracy at the individual level. © 2005 Nature Publishing Group. All rights reserved.