[Objective] In order to provide useful technical support for performance evaluation of irrigation and fertilization in jujube plantations, a jujube shoot biomass estimation method without damaging trees was explored. [Methods] Based on a large number of measurements, both linear models and non-linear models were used to estimate the biomass of branches, hangings, fruits and leaves for jujube. In addition, these models were evaluated and compared according to the correlation coefficient(R²), fit index(W), Standard error(SEE), coefficient of variation(C_V) and forecast accuracy(p). [Results] The linear models exhibited a superior performance for branch biomass modeling with R² ranged from 0.980 to 0.984, p ranged from 0.909 to 0.926. While, non-linear models were more suitable for hanging and leaf biomass modeling(R² was between 0.880 and 0.943, p ranged from 0.897 to 0.976). The optimal fruit biomass model is the dual non-linear model (the R² is 0.999 and p is 0.998). The ratio of fresh and dry weight for branch, hanging and leaf was 2.085, 2.675 and 2.854, respectively. A good linear relationship was showed between the length of a single jujube hanging and the number of leaves on it (R² reaches 0.88).[Conclusion] In this reserch, the simulation precision of the jujube tree biomass model was higher.