In the ion selective electrode (ISE) based soil nitratenitrogen (NO-3N) detection, coexisting chloridion (Cl-), changing of temperature and soil moisture are the primary interference factors. Aiming at improving the accuracy and timeliness of NO-3N detection on site, measurement models for eliminating interference factor’s effects were studied. First, theoretical analysis of modeling and fresh field soil immediate detection was carried on. Modeling theoretical analysis was based on ISE detection principle and multivariate calibration in chemometrics. And to improve timeliness of soil NO-3N detection on site, how to detect fresh field soil immediately and calculate NO-3N content were discussed. Second, two groups of experiments were conducted. One was ISE potential signals collection of mixed solutions with different concentrations (NO-3 and Cl-) at various temperatures measured by detection instrument. According to modeling dataset of ISE potential signals, interference factor’s effects were discussed and three kinds of measurement models were built, including unitary linear regression model of 25℃(base model), multivariate linear regression model of 25℃(simultaneous detection of NO-3 and Cl-), and multivariate temperature correction model of 5~30℃(simultaneous detection of NO-3 and Cl- at changing temperature). And the accuracy of three models were verified and compared according to validation dataset. The other one was NO-3N content detection of fresh field soil measured by corrected and uncorrected NO-3N content conservation equation. The results showed that when Cl- existed and environment temperature changed, multivariate linear regression model of 25℃ was the best (relative error of NO-3 and Cl-  was respectively less than -8.37% and -12.03%). Besides, corrected soil NO-3N content conservation equation can reduce the error caused by soil moisture.