A　new　Cloud-based　Trust　Awareness　and　Interaction　Model　(CTAIM)　is　proposed　for　trust　evaluation　based　content　filtering　in　social　interactive　data.　The　research　is　based　on　diversified　information　from　social　interactive　data　to　analyse　people＇s　intention　and　evaluate　people＇s　trust.　The　proposed　model　is　composed　of　Bayesian　content　filtering　algorithm　and　Bayesian　inference　algorithm　in　Dirichlet　distribution,　and　it＇s　capable　to　provide　3rd　party　trustworthiness　evaluation　according　to　node　behavior　and　interaction　history　with　high-efficiency,　security,　and　neutrality.　Additionally,　MapReduce-based　computing　and　HBase　storage　framework　are　implemented　for　parallel　computing　among　mass　interactive　data.　©　2014　IEEE.