We investigate the ocular dimensions and shape by using Lenstar900 (LS900), A-scan ultrasonography, and Magnetic Resonance Imaging (MRI) in highly myopic Macaca fascicularis. The ocular dimensions data of LS900, A-scan ultrasonography and MRI was assessed from 8 eyes (4 adult male cynomolgus macaque) with extremely high myopia (≤-1000DS) and compared by means of coefficients of concordance and 95% limits of agreement. Multiple regression analysis was performed to explore the associations between ocular biometry, volume, refraction and inter-instrument discrepancies. Test-retest reliability of three measurements of ocular parameters at two time points was almost equal (intraclass correlation= 0.831 to 1.000). The parallel-forms reliability of three measurements was strong for vitreous chamber depth (VCD) (coefficient of concordance = 0.919 to 0.981), moderate for axial length (AL) (coefficient of concordance = 0.486 to 0.981), and weak for anterior chamber depth (ACD) (coefficient of concordance = 0.267 to 0.621) and lens thickness (LT) (coefficient of concordance = 0.035 and 0.631). The LS900 and MRI systematically underestimated the ACD and LT comparing to A-scan ultrasound (P<0.05). Notably, the average AL on LS900 displayed a significant correlation with those on MRI (r=0.978, P<0.001) and A-scan ultrasonography (r=0.990, P<0.001). Almost 4/5 eyeballs were prolate. The mean eyeball volume positively correlated with AL (r=0.782, P=0.022), the width (r=0.945, P=0.000), and the length (r=0.782, P=0.022) of eyeball, while negatively correlated with SER (r=-0.901, P=0.000). In conclusion, there was a high inter-instrument concordance for VCD with LS900, A-scan ultrasonography and MRI, while ACD and LT were underestimated with LS900 compared to A-scan ultrasonography, and the LS900 and A-scan ultrasonography could reliably measure the AL. MRI further revealed an equatorial globe shape in extremely myopic non-human primates.