Space-Terrestrial　Integrated　Networks　(STINs)　and　ubiquitous　artificial　intelligence　are　considered　as　important　infrastructures　of　6G　networks.　Due　to　the　long　transmission　distance　and　high　moving　speed　of　on-orbit　satellites,　STINs　face　the　challenges　of　unreliable　data　transmission,　untimely　data　processing,　and　low　utilization　of　satellite　resources.　Tiansuan　constellation　implements　the　edge　computing　platform,　which　builds　the　foundation　of　edge　intelligence　for　STINs.　This　article　proposes　a　cloud-edge　aggregated　artificial　intelligent　architecture　for　STINs.　With　the　assistance　of　Tiansuan　constellation,　an　implementation　strategy　of　the　intelligent　network　architecture　is　proposed.　Furthermore,　three　use　cases　are　performed　and　discussed　to　achieve　intelligent　data　transmission,　intelligent　remote　sensing,　and　intelligent　network　slicing.　Simulation　results　demonstrate　that　the　proposed　architecture　can　fully　utilize　the　resource　of　STINs　and　improve　the　overall　network　throughput　effectively.