An aggregate model that accurately quantifies buildings' thermal flexibility is an essential interface for energy system operation. However

privacy concerns and measurement inaccuracies make obtaining an accurate aggregate model challenging. To address this

we propose a privacy-preserved robust parameter estimation approach to get the aggregate model of buildings. The proposed method ensures the accurate modeling of the aggregate thermal dynamics of buildings against outliers in measurements

without exposing any sensitive information from end users. The effectiveness of the proposed method is validated through numerical tests.