Background: We prospectively evaluated the incidence and possible factors causing intramuscular injection during lumbar sympathetic ganglion block and compared the multiple needle technique to the single technique to obtain a profound and complete block effect. Methods: Among 83 patients, 58 patients (group A, n = 27, multiple needle technique and group B, n = 31, single needle technique) were reevaluated for the changes of skin temperature (Ts) and mean segment of longitudinal contrast spread. After injecting the contrast agent, the incidence of psoas muscle injection and the change of Ts was compared between two groups. Results: The incidence of psoas muscle injection was 21.3% (46/216) and it was associated with the level of injection (L2) significantly (${chi}_2$ = 14.773, P = 0.001). $DT^{post}$ (postblock temperature difference between ipsilateral and contralateral great toe, $4.6{pm}2.8^{circ}C$, $1.8{pm}1.6^{circ}C$, P < 0.001 for group A and B) and $DT^{net}$ ($DT^{post}$ - $DT^{pre}$, $3.9{pm}2.7^{circ}C$, $1.5{pm}1.5^{circ}C$, P < 0.001 for group A and B) was significantly higher in group A. The mean segment of longitudinal contrast spread was $8.1{pm}0.9$ for group A and $3.2{pm}1.6$ for group B (P < 0.001). Conclusions: The LSGB at the L2 level showed the lowest incidence of psoas muscle injection of contrast. Multiple needle approach showed more significant increase of $DT^{net}$ and $DT^{post}$.