Introduction: dental caries is a public health problem affecting a large percentage of the population. The carious process is highly variable and has periods of progression that alternate with periods of stability of the damaged tissue. There are different techniques to diagnose dental caries, including clinical and radiographic evaluation. Objective: the objective of this study was to establish correlation between the clinical and radiographic caries diagnosis suggested by ICCMSTM, in deciduous and permanent molars of a school population. Methods: descriptive study evaluating a sample for convenience of 1174 proximal and occlusal tooth surfaces of permanent and deciduous molars, taken from the database of 35 outpatients treated at the school of dentistry, who were clinically and radiographically evaluated for caries as recommended by the ICCMSTM based on bitewing x-rays. Results: the clinical and radiographic diagnosis was correlated in 1174 proximal and occlusal surfaces, with 0.41 Spearman's rank correlation coefficient (p < 0,05). The findings suggest that 95.6% of teeth diagnosed as healthy coincided with the clinical and radiographic results; in early mild stages, there was coincidence in only 8.16% and 6.4% respectively. Conclusions: there is low correlation between the clinical diagnosis of caries and the radiographic examination, in relation to ICCMSTM standards.