Abstract Hypokalaemia is common in hypomagnesaemic patients, but the in vivo mechanisms have not been determined. Along the distal nephron, lower intracellular Mg 2+ has been proposed to release the Mg 2+ ‐mediated inhibition of renal outer medullary K + (ROMK) channels, increasing urinary K + excretion. Higher activity of the epithelial Na + channel (ENaC), which provides the driving force for K + secretion via ROMK, may also be required. We tested the hypothesis that hypomagnesaemia‐induced hypokalaemia is associated with higher activities of both ENaC and ROMK. C57BL/6J mice were fed normal (NL), low Mg 2+ (LM), low Na + (LS) or low Na + /Mg 2+ (LS/LM) diets. Kidneys and blood were harvested for western blotting and measurement of plasma [K + ]. ROMK activity was measured by patch‐clamping of split‐open distal convoluted tubule 2 (DCT2)/connecting tubule (CNT) from mice on NL or LS/LM diets. Plasma [K + ] was significantly lower in mice on LS/LM diet. Compared with mice on NL diet, abundances of cleaved α‐ and γ‐ENaC, which correlate with ENaC activity, were higher in mice on LS diet, and lower in mice on LM diet but not on LS/LM diet, suggesting Na + and Mg 2+ restriction counteract each other. A lower natriuretic response to acute amiloride administration confirmed lower ENaC activity in mice on LM diet. ROMK activity along DCT2/CNT was higher in mice on LS/LM but not on LM diet compared with mice on NL diet. Together, our data suggest that Mg 2+ restriction inhibits ENaC activity, and both higher ENaC activity and disinhibition of ROMK are required for the development of hypokalaemia in the context of hypomagnesaemia. image Key points In patch‐clamping experiments, lower intracellular Mg 2+ releases the Mg 2+ ‐mediated inhibition of renal outer medullary K + (ROMK) channels, increasing urinary K + excretion, but hypomagnesaemia does not always lead to hypokalaemia in several genetic disorders and animal experiments. Epithelial sodium channel (ENaC)‐mediated Na + entry provides the driving force for K + secretion via ROMK, but how hypomagnesaemia affects ENaC activity remains unknown. We found that ENaC activity is lower following dietary Mg 2+ restriction but is preserved with combined Mg 2+ /Na + restriction, stimulating K + secretion via ROMK. Our findings probably explain why Mg 2+ restriction alone cannot cause hypokalaemia in vivo and provide new insights into the mechanisms of hypomagnesaemia‐induced hypokalaemia.