This study investigated the effects of a number sense intervention in 5-year-old Italian children (n = 96) from different socioeconomic status (SES) backgrounds. We investigated whether a short-term intervention could enable low SES children to reach the same level of their middle-upper SES peers in number sense; in addition, we compared the efficacy of the intervention for low SES children at two levels of intensity. Four groups of children were compared as follows: 2 low SES groups that received 8 or 20 intervention sessions, 1 middle-upper SES group that received eight intervention sessions, and 1 middle-upper SES control group. The results show that at the baseline, the middle-upper SES children obtained higher scores than the low SES children in three of the four scales of the Battery of Numerical Intelligence (BIN). Regarding the effectiveness of the intervention, all intervention groups showed a higher change in the counting subscale score than the control group. The low SES group that received the high-intensity intervention showed a greater change in the BIN total score than all other groups.