Childhood acute lymphoblastic leukemia (ALL) peaks around age 2–4, and in utero genetic epigenetic mother-fetus crosstalk might tune ALL onset during childhood life. Folate genes variably interact with vitamin status on ALL risk and prognosis. We investigated DHFR and MTHFR gene variants in 235 ALL children and their mothers to disclose their role in determining ALL onset age and survival. Pyrosequence of DHFR 19bp ins/del (rs70991108; W/D), MTHFR C677T (rs1801133; C>T), and MTHFR A1298C (rs1801131; A>C) was assessed in children and in 72% of mothers for dyad-analysis comparison. DHFR DD-children had delayed ALL onset compared to WW-children (7.5 ± 4.8 vs. 5.2 ± 3.7 years; P = 0.002) as well as MTHFR 1298 CC-children compared to AA-children (8.03 ± 4.8 vs. 5.78 ± 4.1 years; P = 0.006), and according to the strong linkage disequilibrium between MTHFR 677 T-allele and 1298C-allele, MTHFR TT-children showed early mean age of onset though not significant. Offspring of MTHFR 677 TT-mothers had earlier ALL onset compared to offspring of 677 CC-mothers (5.4 ± 3.3 vs. 7 ± 5.3 years; P = 0.017). DHFR/MTHFR 677 polymorphism combination influenced onset age by comparing DD/CC vs. WW/TT children (8.1 ± 5.7 vs. 4.7 ± 2.1 years; P = 0.017). Moreover, mother-child genotype combination gave 5.5-years delayed onset age in favor of DD-offspring of 677 CC-mothers vs. WW-offspring of 677 TT-mothers, and it was further confirmed including any D-carrier children and any 677 T-carrier mothers (P = 0.00052). Correction for multiple comparisons maintained statistical significance for DHFR ins/del and MTHFR A1298C polymorphisms. Unexpectedly, among the very-early onset group (<2.89 years; 25th), DD-genotype inversely clustered in children and mothers (4.8% vs. 23.8% respectively), and accordingly ALL offspring of homozygous DD-mothers had increased risk to have early-onset (adjusted OR (odds ratio) = 3.08; 1.1–8.6; P = 0.03). The opposite effect DHFR promoter variant has in tuning ALL onset-time depending on who is the carrier (i.e., mother or child) might suggest a parent-origin-effect of the D-allele or a two-faced epigenetic role driven by unbalanced folate isoform availability during the in-utero leukemogenesis responsible for the wide postnatal childhood ALL latency.