Work zone environments are complex, and adequate capacity estimation methods must consider more factors than regular freeway sections. In order to avoid the undue influences on the traffic capacity calculation process and evaluate the impact of work zones on operating efficiency, this work studies the capacity of two-way four-lane freeway work zones using three methods: saturation flow method, the Highway Capacity Manual 2016 method, and a simulation approach. Peak hour traffic volumes of each section collected on Xi’an-Lintong freeway work zone are used to compare the results of the three methods. The saturation flow rate (SRF) method requires estimating average headway values (in this study obtained by fitting an improved Weibull distribution to field data), which leads to a capacity of 1,535 passenger cars per hour per lane (pcphpl). The corresponding capacity values using HCM and simulation are 1,611 pcphpl and 1,491 pcphpl, respectively. The results of the three calculation methods show that the SFR method is also an attractive alternative for work zone capacity estimation. In addition, the improved Weibull distribution used to model headways was observed to provide a very good fit for the field data collected in this study; the corresponding average headways are expected to be reliable and applicable to similar work zones in China.