The invention discloses a high-density packaged element positioning method based on speeded up robust features (SURFs), which comprises the following steps: (1), coarse positioning: coarse transformation parameters (m, n and Beta) are obtained by adopting a SURF registration method; (2) fine positioning: an image I (x, y) to be registered is inversely transformed by utilizing the coarse transformation parameters (m, n and Beta) to obtain an image g (x and y), the minimum energy equation E of the image g (x and y) and a template image f (x and y) is calculated, each partial derivative of the E is solved and enabled to be zero, and transformation parameters (a, b and Theta) are solved; m' is enabled to be equal to m plus a, n' is enabled to be equal to n plus b, and Beta' is enabled to be equal to Beta plus Theta; the energy equation E under a transformation relation (m', n' and Beta') is calculated; whether the E is smaller than a set value or not is judged, and if no, next iteration is conducted; and if yes, the iteration process is finished, m' is taken as the final translation parameter in an x direction, n' is taken as the final translation parameter in a y direction, and Beta' is taken as the final rotation angle. Compared with the prior art, the high-density packaged element positioning method realizes high-precision sub-pixel positioning and has great robustness on illumination transformation and noise.