The　achievement　of　well-performing　pressure　sensors　with　low　pressure　detection,　high　sensitivity,　large-scale　integration,　and　effective　analysis　of　the　subsequent　data　remains　a　major　challenge　in　the　development　of　flexible　piezoresistive　sensors.　In　this　study,　a　simple　and　extendable　sensor　preparation　strategy　was　proposed　to　fabricate　flexible　sensors　on　the　basis　of　multiwalled　carbon　nanotube/polydimethylsiloxane　(MWCNT/PDMS)　composites.　A　dispersant　of　tetrahydrofuran　(THF)　was　added　to　solve　the　agglomeration　of　MWCNTs　in　PDMS,　and　the　resistance　of　the　obtained　MWCNT/PDMS　conductive　unit　with　7.5　wt.%　MWCNTs　were　as　low　as　180　omega/hemisphere.　Sensitivity　(0.004　kPa(-1)),　excellent　response　stability,　fast　response　time　(36　ms),　and　excellent　electromechanical　properties　were　demonstrated　within　the　pressure　range　from　0　to　100　kPa.　A　large-area　flexible　sensor　with　8　x　10　pixels　was　successfully　adopted　to　detect　the　pressure　distribution　on　the　human　back　and　to　verify　its　applicability.　Combining　the　sensor　array　with　deep　learning,　inclination　of　human　sitting　was　easily　recognized　with　high　accuracy,　indicating　that　the　combined　technology　can　be　used　to　guide　ergonomic　design.