By introducing image recognition and micro-current testing, jet behavior research was conducted, in which the real-time recognition of ejection mode was realized. To study the factors influencing ejection modes and the current variation trends under different modes, an Electrohydrodynamic Direct-Write (EDW) system with functions of current detection and ejection mode recognition was firstly built. Then a program was developed to recognize the jet modes. As the voltage applied to the metal tip increased, four jet ejection modes in EDW occurred: droplet ejection mode, Taylor cone ejection mode, retractive ejection mode and forked ejection mode. In this work, the corresponding relationship between the ejection modes and the effect on fiber deposition as well as current was studied. The real-time identification of ejection mode and detection of electrospinning current was realized. The results in this paper are contributed to enhancing the ejection stability, providing a good technical basis to produce continuous uniform nanofibers controllably. (c) 2018 Author(s).