<i>Objective:</i> The aim of this trial was to evaluate the correlation between amniotic fluid index (AFI) and the real amniotic volume in mid-trimester pregnancies. <i>Methods:</i> Eight women with mid-trimester anhydramnios pregnancies were included. Patients gave their informed consent. Those with premature rupture of membranes were excluded. Amnioinfusion was performed by instilling up to 400 ml of saline solution in 100 ml aliquots. With the patient in supine position, the AFI was determined by the sum of measurements of the deepest vertical pools in each of the 4 quadrants of the maternal uterus, evaluating the AFI after every 100 ml of infused solution. Statistical analysis was performed by analysis of variance, determination coefficient, linear correlation and t test. <i>Results:</i> We realized 29 AFI measurements after amnioinfusions of 100, 200, 300 and 400 ml of saline solution. The AFI ranges that corresponded to the infused volume were: after 100 ml = 5–11 cm (median 6.5 cm); after 200 ml = 9–12.7 cm (median 10.3 cm); after 300 ml = 10–17 cm (median 13.5 cm), and after 400 ml = 12–16 cm (median 15 cm). The correlation between the AFI and the volume infused was r = 0.81 (p < 0.0001), the variance was r² = 0.65, suggesting that 65% of the variation in AFI measurements is directly accounted for by the amniotic fluid volume and 35% of the AFI measurements is accounted for by factors other than amniotic volume. <i>Conclusions:</i> We obtained a strong correlation between AFI and amnioinfusion volume (median of 6.5, 10.3, 13.5 and 15 cm after amniotic infusion of 100, 200, 300 and 400 ml of saline solution, respectively). The variance (0.65) suggests that 35% of AFI measurements is accounted for by factors other than amniotic fluid volume.