The invention discloses a Barkhausen noise lossless measuring method and system for a ferromagnetic material local stress and strain field. A hard disc magnetic head is used for detecting a Barkhausennoise signal in a cyclic magnetization process of a ferromagnetic material, multiple extracted magnetic parameters are input to a pre-established force-magnetic parameter mapping relation neural network model, and a size of the stress or the strain is quantitatively forecasted. The hard disc magnetic head of which a magnetic core terminal air gap is smaller than 1 um is carried by a three-axis moving platform and used for scanning along a local area range of the measured material. A Barkhausen noise detecting device is capable of synchronously detecting the Barkhausen noise signal. The neuralnetwork model is capable of forecasting the size of the stress or the strain of a scanning point, and acquiring the stress field or the strain field within the local area range of the material afterscanning. So a relation of the Barkhausen noise signal in the ferromagnetic material and the stress and strain is calibrated, and rapid and lossless quantitative detection is performed on the local stress and strain field.