This study investigates the characteristics and application of the optically-stimulated luminescence dosimeter (OSLD) in cobalt-60 high dose rate (HDR) brachytherapy, and compares the results with the dosage produced by the treatment planning system (TPS). The OSLD characteristics comprised linearity, reproducibility, angular dependence, depth dependence, signal depletion, bleaching rate and cumulative dose measurement. A phantom verification exercise was also conducted using the Farmer ionisation chamber and in vivo diodes. The OSLD signal indicated a supralinear response (R2 = 0.9998). It exhibited a depth-independent trend after a steep dose gradient region. The signal depletion per readout was negligible (0.02%), with expected deviation for angular dependence due to off-axis sensitive volume, ranging from 1 to 16%. The residual signal of the OSLDs after 1 day bleached was within 1.5%. The accumulated and bleached OSLD signals had a standard deviation of ± 0.78 and ± 0.18 Gy, respectively. The TPS was found to underestimate the measured doses with deviations of 5% in OSLD, 17% in the Farmer ionisation chamber, and 7 and 8% for bladder and rectal diode probes. Discrepancies can be due to the positional uncertainty in the high-dose gradient. This demonstrates a slight displacement of the organ at risk near the steep dose gradient region will result in a large dose uncertainty. This justifies the importance of in vivo measurements in cobalt-60 HDR brachytherapy.