Purpose. To compare long-term variability of retinal nerve fiber layer thickness (RNFLT) measurements made with the RTVue-100 Fourier-domain optical coherence tomograph (RTVue-OCT) and scanning laser polarimetry with variable (GDx-VCC) and enhanced (GDx-ECC) corneal compensation. Methods. One eye each of 110 Caucasian patients (17 healthy, 20 ocular hypertensive, 22 preperimetric, and 51 perimetric glaucoma eyes, of which 10 showed visual field progression) were imaged prospectively at 6-month intervals for 1.5 to 3 years. Results. Median change of Octopus visual field mean defect was -0.300 dB/y for the controls, -0.120 dB/y for perimetric glaucoma group, and 1.231 dB/y for the 10 functionally progressing perimetric glaucoma eyes. Relative (%) variance of all RNFLT parameters measured with RTVue-OCT was significantly (p<<0.001) smaller than that with both of the GDx methods. Relative RNFLT progression slopes (% change/y) did not differ significantly between the methods. Relative variance of the RNFLT parameters did not differ between the control and perimetric glaucoma group with RTVue-OCT, but for most parameters it was consistently higher for perimetric glaucoma with both GDx methods (p=0.01). Relative RNFLT progression slopes, however, did not separate the control and perimetric glaucoma eyes with either method. Conclusions. Long-term RNFLT measurements are less variable with the RTVue-OCT than with GDx-VCC and GDx-ECC. Long-term measurement variability is higher in perimetric glaucoma than in healthy eyes with both GDx methods, but no similar between-group difference is seen with RTvue-OCT. None of the tested methods separated the rate of progression between healthy and well-controlled perimetric glaucoma eyes in 1.5 to 3 years follow-up.