Online　social　media　provides　people　with　a　platform　to　express　their　emotions　anonymously.　Social　media　has　been　identified　as　an　important　data　source　for　suicide　prevention　related　to　emotional　problems　in　China.　Almost　Three　million　messages　were　published　by　450,000　users　in　a　particular　Chinese　social　media　data　base.　This　study　aims　to　develop　a　Crisis　Balance　Analysis　Model　based　on　concepts　of　＇balancing　factors＇as　described　by　Aguilera.　Through　interactions　with　psychological　experts,　deep　learning　architecture　that　was　built　and　refined.　Three　annotation　levels　free　annotations　(zero　cost),　easy　annotations　(by　psychology　students),　and　hard　annotations　(by　psychology　experts)　were　used.　Our　Model　was　evaluated　accordingly　and　showed　that　its　performance　at　each　level　was　promising.　Finally,　suicide　risks,　cognitive　distortions　and　interpersonal　problems　could　be　identified　for　messages　from　social　media　users　using　this　model,　which　providing　basis　for　proactive　crisis　intervention.　©　2021　ACM.