Skills　possessed　by　human　welders　typically　require　a　long　time　to　develop.　Especially,　maintaining　the　torch　to　travel　in　desired　speed　is　challenging.　In　this　paper,　a　feedback　control　system　is　designed　and　implemented　to　assist　the　welder　to　adjust　the　torch　movement　for　the　desired　speed　in　manual　gas　tungsten　arc　welding　(GTAW)　process.　To　this　end,　an　innovative　helmet　based　manual　welding　platform　is　proposed　and　developed.　In　this　system,　vibrators　are　installed　on　the　helmet　to　generate　vibration　sounds　to　instruct　the　welder　to　speed　or　slow　down　the　torch　movement.　The　torch　movement　is　monitored　by　a　leap　motion　sensor.　The　torch　speed　is　used　as　the　feedback　for　the　control　algorithm　to　determine　how　to　change　the　vibrations.　To　design　the　control　algorithm,　dynamic　experiments　are　conducted　to　correlate　the　arm　movement　(torch　speed)　to　the　vibration　control　signal.　Linear　model　is　firstly　identified　using　standard　least　squares　method,　and　the　model　is　analyzed.　A　nonlinear　Adaptive　Neuro-Fuzzy　Inference　System　(ANFIS)　model　is　then　proposed　to　improve　the　modeling　performance.　The　resultant　nonlinear　ANFIS　model　can　estimate　the　welder＇s　response　on　the　welding　speed　with　acceptable　accuracy.　Based　on　the　response　model,　a　PID　control　algorithm　has　been　designed　and　implemented　to　control　the　welder　arm　movement　for　desired　torch　speed.　Experiments　verified　the　effectiveness　of　the　system　for　the　desired　speed　with　acceptable　accuracy.