The association of inadequate dietary sodium intake with bone mineral density (BMD) and the risk of osteoporosis is controversial. To find the association between low sodium diet and the risk of incipient osteoporosis, we performed a population-based cross-sectional analysis using Tanaka method for estimation 24-h urinary sodium excretion (e24hUNaETanaka) as a candidate indicator of sodium intake. We identified 3869 participants without osteoporosis and classified them into quartiles according to their value of e24hUNaETanaka. BMD was measured to find participants at risk of osteoporosis. Lower e24hUNaETanaka was related to decreasing BMD of the distal radius. Multiple Cox-proportional hazard models demonstrated that e24hUNaETanaka had an inverse association with the risk of osteoporosis (adjusted HR = 0.859, 95% CI = 0.751–0.982) and survival analysis revealed that the lowest quartile group had poor osteoporosis-free survival (PLog-rank < 0.0001). Furthermore, our restricted cubic spline analysis revealed that the relationship between e24hUNaETanaka and HR of osteoporosis was negative curvilinear in males and postmenopausal females and positive linear in premenopausal females. Our findings suggest that lower sodium intake was a significant predictor of incipient osteoporosis and there was wide variation in this relationship according to sex and female hormone status.