Membrane distillation (MD) is a promising desalination technology that combines traditional distillation processes with membrane separation technology, but when the feed solution contains amphiphilic substances such as surfactant, the hydrophobic membrane is prone to being wetting, which hinders its widespread application. In this study, a novel Janus membrane was developed by embedding Zr-MOF and silver nanowires into the modified graphene oxide composite membrane to improve its separation performance, and the effect of high-concentration SDS on the membrane performance was studied. The experimental results showed that the conductive membrane exhibited excellent performance when powered on for 5 min every 2.5 h under 2 V voltage. The flux stabilized at 19.6 L/(m2h) with a desalination rate up to 99.9 % and no obvious signs of membrane wetting. It was found that the effect of electric field caused a negative charge on the membrane surface, forming a charge repulsion force with the SDS. Simultaneously, intermittent electrical flotation can effectively remove SDS that has accumulated on the membrane surface, thus achieving efficient and stable operation of MD system. Moreover, a machine learning approach with Catboost was applied to predict and unveil the complex effects on membrane performance, providing a deep insight to treat wastewater containing surfactant.