Japan's super-aged population faces significant health challenges, with physical pre-frailty and frailty affecting nearly half of older adults alongside high cardio-metabolic multimorbidity rates. Environmental stressors like atmospheric
pollution and poor infrastructure accessibility further accelerate physical decline. Grounded in Japan's "Society 5.0" framework, the GPONE project introduces a Public Health Support SaaS platform that integrates Earth Observation satellite remote sensing with official e-Stat census data and MHLW National Database clinical claims. The platform utilises a multi-modal AI pipeline combining Vision-Language Models for spatial satellite rasters and Retrieval-Augmented Generation for clinical data to feed a fine-tuned medical Large Language Model. This system powers an interactive Public Health Advisory Dashboard that allows municipal authorities and insurers to query complex environment-disease correlations using plain natural language. By enabling targeted early interventions, the platform helps prevent care dependency and optimises public health resource allocation.