This study was initiated to evaluate o-cresol and benzylmercapturic acid in urine in comparison with other biomarkers, as tools to estimate the intensity of occupational exposure to toluene at low levels. In total, 108 solvent exposed workers (engaged in tape production) and 17 non-exposed controls (all men) participated in the study. The surveys were conducted in the second half of working weeks. Diffusive sampling was conducted to measure 8-h time-weighted average intensity of occupational exposure to toluene. Blood and urine samples were collected at the end of a working shift. Blood samples were subjected to analysis for toluene (Tol-B), and urine samples were analyzed for benzyl alcohol (BeOH-U), benzylmercapturic acid (BMA-U), o-cresol (o-CR-U), hippuric acid (HA-U) and toluene (Tol-U) by the methods previously described. The toluene concentrations in workplaces were low in general, with a geometric mean (GM) and the maximum concentration of 1.9 ppm and 8.8 ppm, respectively. The statistical analyses of the six biomarkers for correlation with air-borne toluene showed that both Tol-B and Tol-U gave a high correlation coefficient of 0.58 to 0.61 (p<0.01), whereas the coefficients for BeOH-U and BMA-U together with HA-U were all low (up to 0.22, depending on the correction for urine density) and statistically insignificant (p>0.10) in most cases. o-CR-U had an intermediary coefficient of 0.20 (p<0.05). Comparison with previous publications disclosed that BeOH-U, BMA-U and HA-U correlate with toluene in air when the exposure is intense (e.g., 50 ppm or above), but no longer proportional to air-borne toluene when the exposure is low, e.g., 2 ppm. Such appeared to be also the case for o-CR-U. In over-all evaluation, the validity of o-CR-U in monitoring occupational exposure to toluene at low levels (e.g., 2 ppm) appear to be limited, and BMA-U is not an appropriate biomarker. BeOH-U and HA-U are also inappropriate for this purpose. Only Tol-B and Tol-U may be employed to estimate toluene exposure at low levels.