Purpose: We develop a new low junction-dose-gradient VMAT planning technique to mitigate the position-induced dose error based on standard linac with true attachment free. Methods: An eight years old girl diagnosed with acute lymphocytic leukemia underwent TBI in 2020. Target volumes and organs at risk were contoured on 2 sets of CT scan done in head-first supine and feet-first supine positions respectively. Two component plans were then created in sequence, firstly a base plan that covered the caudal half of the target volume with shallow dose gradient toward the cranial end, secondly a plan that covered the cranial half with matching dose gradient toward the caudal end. The final product was a composite plan that is dosimetrically matched at the junction region of the 2 component plans. Results: Average time for on-board image position verification per fraction was 40.3 minutes. Average beam on time per fraction was 37.2 minutes. Average time for changing from head first to feet first treatment position and on board imaging verification was 18.4 minutes. Average mean lung dose was 9.89 Gy, Maximum lens dose was 7.6 Gy. Mean dose and maximum dose of planning target volume (PTV) were 12.17 Gy and 13.65 Gy respectively. 98.23% of PTV was covered by 90 % of the prescribed dose. Mean dose and V10.8 of PTV are only slightly different (0.49% - 1.89% and 0.26% - 1.04% respectively) even with an error of 5 - 20mm longitudinal misalignments. Gamma passing rate (3mm/ 3% Gamma criteria) are between 93.5% and 100%. Real-time dose monitoring showed an overall deviation of -3.9% ± 5.51%. Conclusions: Low junction-dose-gradient VMAT based on standard linac with true attachment free for total body irradiation is feasible, accurate, and reliable. It is worthy of clinical application.