The rapid development of highway traffic has gradually deteriorated the acoustic environment along highways. Sonic crystal theory provides new ideas for traffic noise control. However, current road noise reduction methods based on sonic crystals are still at the theoretical analysis stage, and the lack of reliable calculation models and field test verifications has restricted the promotion and application of sonic crystal acoustic barriers (SCABs). Based on the acoustic performance design of SCABs, in this study, a field test was conducted to study the noise reduction characteristics of SCABs; the corresponding 3D finite element method (FEM) model was established and validated; and three types of optimization measures for noise reduction performance were explored. The main conclusions were as follows: Compared with traditional acoustic barriers along highways, SCABs are more targeted for noise reduction, and the noise reduction performance in the band gap is improved by 0.3-8.8 dB(A). Moreover, as the distance increases behind the acoustic barriers, the noise reduction effect decreases more slowly. The 3D FEM model can accurately describe the spatial characteristics of the sound field and shows agreement with the field test results, with an error of less than 3 dB(A). Compared with the 2D FEM model, the 3D FEM model can accurately calculate the noise reduction performance of an SCAB with resonant elements. Additionally, the gradient combination of scatterers, the filling of porous sound-absorbing materials and the use of microperforated plates can effectively improve the SCAB noise reduction performance. The results provide scientific and reliable experimental support and design guidance for the popularization and application of SCABs in traffic noise control.