People are inundated with messages about foods' healthiness/unhealthiness and capacity for weight gain/loss. This study examined how people evaluate foods for 'healthiness/unhealthiness' or 'capacity for weight gain/loss' and whether these evaluations influence 'caloric' estimation accuracy. Fifty-five participants in a behavioral weight loss program rated eight 'healthy' and eight 'unhealthy' foods on perceived 'healthiness/unhealthiness', 'capacity to influence weight loss or gain' and 'caloric' content. Open-ended questions inquired why a food was 'healthy/unhealthy' or would 'contribute to weight gain/loss'. Open-ended questions were coded into 29 independent food categories (e.g. high fat). While similar foods were often judged as 'healthy' and 'contributing to weight loss' or 'unhealthy' and 'contributing to weight gain', participants provided different explanations for their 'health' and 'weight' ratings. Healthy/weight loss foods were systematically underestimated in 'calories', while unhealthy/weight gain foods were systematically overestimated in 'calories'. Errors in 'caloric' estimation decreased from pre- to post-weight loss treatment, but were not associated with weight loss. Higher baseline BMI was associated with greater inaccuracy in 'caloric' estimation. Fat content was the most common category used for evaluating foods. Participant evaluations of healthy/weight loss foods or unhealthy/weight gain foods were systematically associated with 'caloric' estimation.