This study aims to forecast the construction cost index (CCI) by utilizing multivariate time series, VAR(Vector Autoregression). To address
the limitations of traditional univariate models, especially in times of economic uncertainty, we identify leading indicators through statistical
analysis. Three key leading indicators—construction order value, business survey index, and producer price index for structural steel—are selected. VAR model is developed and optimized via AIC, with its performance evaluated using Walk-Forward cross-validation over a 25-year dataset. The proposed multivariate framework provides a robust approach for forecasting CCI by incorporating the dynamic influence
of leading market indicators.