Objective To compare the effect of interscalene brachial plexus block using three kinds of ultrasound-guided injection methods. Methods Ninety ASA Ⅰ-Ⅱpatients scheduled for surgery of the shoulder or the upper extremity were randomly divided into three groups(n=30), S group: ultrasound-guided and single-point-injection group; M group: ultrasound-guided and three-point-injection group; R group: ultrasound-guided multiple injection group using as little local anesthetic as assessed by sonography that the nerve was soaked by local anesthetic. A total dose of 30 ml ropivocaine was injected in the S and M group. In the R group, if 0.5% the stems of brachial plexus were coated to show "donut sign", then we stopped injecting local anesthetic. Results The average onset time of sensory blockade on the area innervated by the medial antebrachial cutaneous nerve and the medial brachial cutaneous nerve in the M (7.6±2.3) min and(6.0±2.2) min and R (8.2±2.6) min and (7.6±2.7) min groups was significantly shorter than that in the S (9.2±3.1) min and (10.6±2.8) min group. The rate of satisfactory sensory blockade on the area innervated by the medial brachial cutaneous nerve in the S(60%) group was significantly lower than in the M(100%) and R (94.3%) groups. The incidence of complications in the M(60%) group was significantly higher than in the S (20%) groups. Conclusions The ultrasound-guided multi-point interscalene brachial plexus nerve block can speed up the onset time of the medial antebrachial cutaneous nerve and the medial brachial cutaneous nerve and improve the success rate of interscalene brachial plexus nerve block. Multi-point injection of local anesthetic can increase the incidence of complications.