It is currently debated whether vitamin D requirements during pregnancy differ from those during non-gravid states. In current analyses, we aimed to determine the best model for the association between PTH and serum 25-hydroxyvitamin D (25(OH)D) and the threshold for circulating 25(OH)D at which serum parathyroid hormone (PTH) is suppressed. This multicenter prospective cross-sectional study was conducted on 227 Iranian pregnant women aged 15–45 years in their third trimester of pregnancy. The locally weighted smoothing scatter plot (LOWESS) was used to determine the curvilinear shape of the 25(OH)D/PTH relationship. Linear and non-linear methods were employed to determine the best fit and cut-point for serum 25(OH)D concentration. The median serum 25(OH)D and corresponding serum PTH concentration were 17⋅26 (13⋅44–23⋅08) ng/ml and 19⋅46 (15⋅08–25⋅04) pg/ml in our study population, respectively. The LOWESS curve suggested a non-linear and monotonic with a negative slope relation between PTH (pg/ml) and serum 25(OH)D (ng/ml). The optimal model for the association between PTH and serum 25(OH)D was a one-term fractional polynomial (FP1) (AIC = 1640⋅463). The FP1 analysis identified the 25(OH)D threshold of 12⋅48 ng/ml at which serum PTH rapidly rose. The expected degree of PTH stimulation seems to have a linear trend as 25(OH)D falls below 40 ng/ml. 25(OH)D (ng/ml) and PTH (pg/ml) had a non-linear and monotonic relationship with a negative slope. Our data suggest that a 25(OH)D threshold of 12⋅48 ng/ml is sufficient for parathyroid hormone suppression, which could be used to screen for deficient individuals.