In cheesemaking, the on-line measurement of clotting time, coagulum firmness and cutting point is useful for process monitoring but has proven difficult to carry out. This work presents a comparison of on-line sensors, more or less capable of allowing the determination of these data. An experiment was undertaken in which six on-line sensors, comprising of optical systems (NIR transmission X2 wavelengths and NIR reflectance), hot-wire, and vibrational systems (torsional and transverse), were used simultaneously to monitor milk coagulation in trials where protein levels and enzyme concentrations were varied and the sensor response characteristics were compared with oscillatory shear measurements, It was found that protein level mainly influenced curd firming rate, while rennet level mainly influenced gel time. Up to three coagulation parameters were derived using each sensor. While all sensors were able to detect the effect of rennet level, which is mainly seen in the gelation (or clotting) time, the sensors varied in their sensitivity to changes in protein level, which mainly affects the curd-firming time. For some sensors, a combination of response parameters was shown by multiple linear regression to be advantageous for predicting coagulum-firming time. By combining response parameters in this way the standard errors of prediction of coagulum cutting point (using oscillatory shear measurements as the reference) were better than 100 s for NIR reflectance, NIR transmission and low-frequency vibration systems as compared with ca.300 s for the hot-wire or torsional vibration systems.