Receiver system delay is embedded in additional secondary factor (ASF) observations and coupled with the propagation term at the observation level, which is a key factor limiting high-accuracy ASF measurement in enhanced Loran (eLoran) systems. Existing system delay calibration methods either can only calibrate the internal delay of the receiver and therefore cannot cover the complete delay chain involved in field measurements, or rely on single-point theoretical compensation, making the estimated system delay susceptible to errors in theoretical ASF values. To address this issue, this paper proposes a system delay calibration method based on multipoint ASF measurements for eLoran receivers. By exploiting the structural characteristic that the system delay acts as a shared constant bias in multipoint ASF observations, whereas the ASF spatial component varies with the propagation path, a joint inversion model of system delay and equivalent ground conductivity is constructed to separate the two in the absence of true ASF values. On this basis, a rapid calibration procedure is further established, including calibration area selection, reference receiver calibration, determination of the ASF spatial reference value at a fixed calibration site, and rapid calibration of other receivers. Experimental results in the Xi’an and Zhoukou test areas show that the optimal estimates of the reference receiver system delay are 79.046 μs and 79.048 μs, respectively, with a difference of only 2 ns. After correction for system delay and conductivity, the mean absolute error (MAE) and root mean square error (RMSE) between the ASF spatial component at the validation points and the theoretical ASF are 0.039 μs and 0.045 μs, respectively, in the Xi’an test area, and 0.076 μs and 0.091 μs, respectively, in the Zhoukou test area. Furthermore, the rapid calibration method is validated in the Zhoukou area. For two ASF measurement systems, one using the reference receiver calibrated through joint inversion and the other using a receiver rapidly calibrated at a fixed calibration site, the mean bias, MAE, and RMSE at the same measurement points are reduced by 98.9%, 85.9%, and 84.4%, respectively, compared with the corresponding values before calibration. The results demonstrate that the proposed method can provide relatively consistent system delay estimates under different propagation conditions and improve the consistency of ASF measurements obtained by different receivers.