Land valuation is one of the duties of the Ministry of Agrarian and Spatial Planning / National Land Agency, the need for information on land values is outlined in the form of a Land Value Zone map. However, in practice, the existing ZNT is not accurate and represents the true market value, so that another method of land valuation is needed, in this case another method is the plot-based land valuation. Plot-based land valuation is carried out by collecting variables that affect the value of land (independent variable) in each plot of land, and the sample in this study is the value of land (dependent variable).
Dependent variables and independent variables are used in estimating land values, how to forming valuation models and estimating land values using the random forest algorithm in the ZNT Model Creator and ZNT Estimator applications. In this research, the process of calculating the multiplier factors using the Analytical Hierarchy Process is carried out by structuring each t-test value or the effect of the independent variables partially.
The best accuracy from the use of Random Forest in land valuation is ZNT Model Creator modeling accuracy of 0.752 with tree depth and tree number parameters given the same treatment, namely given values of 2.3 and 4. The modeling is used in estimating soil values using ZNT Estimator with size The accuracy of the estimated land value is the Coefficient Correlation 0.898770667 which means that the correlation between the estimated value and the actual value is very strong, Root Mean Square Error (RMSE) with a value of 631341.1825, Mean Absolute Deviation (MAD) with a value of 501462.2632, Mean Absolute Percentage Error (MAPE) with a value of 26.8417142 which means that the modeling contains 26.84% errors.

Keywords: Land Valuation, Analytical Hierarchy Process, Random Forest