The study examined the association between fundus tissue density (FTD) and subfoveal choroidal thickness (SFCT) in 6159 eyes of children aged 2-6 years with different refractive conditions: hyperopia, pre-myopia, low myopia, and moderate to high myopia. FTD was assessed using artificial intelligence-based analysis of fundus color photographs and SFCT optical coherence tomography. SFCT significantly decreased with increasing severity of myopia (p < 0.001), while FTD increased accordingly (p < 0.001). Significant negative correlations were observed between FTD and SFCT in the pre-myopia (r = -0.183, p = 0.019), low myopia (r = -0.335, p < 0.001) and moderate to high myopia (r = -0.222, p = 0.0008) groups. A non-linear threshold effect was identified: at SFCT below 148.90 μm, FTD increased significantly, while this threshold decreased with higher myopia severity. Central subfield thickness (CST) was not linearly associated with refraction but was increased in moderate to high myopia. AI-based FTD can serve as a non-invasive biomarker for monitoring the progression of myopia in children.