Europe is a highly diversified market, which is driven by European and national interests. Both elements reflect back to the property markets, which react accordingly. Furthermore, the behaviour of market participants is influenced by different irrational elements. All these factors contribute to a specific sentiment within the real estate market. Sentiment is known to play a significant role in asset performances and is used to measure the mood in the market. However, it is an extremely difficult aspect to quantify, especially for real estate assets which are often lumpy investments, infrequently transacted with severe information asymmetry. In the lack of a suitable commercial property sentiment measure we constructed six different indicators which are based on different combinations of sentiment proxies and macroeconomic factors. These sentiment indicators and their proxies are based on the basic idea and method of Baker and Wurgler (2006), who stated that each of the imperfect sentiment proxies includes, at least to a certain share, some pure sentiment. The authors have used an orthogonalization method which regresses the sentiment proxies against macro-economic variables to remove blurring elements. The residuals of these regressions enter a Principal Component Analysis (PCA) for the construction of the sentiment indicators. Using this method a set of different sentiment indicators has been constructed. In addition, a sentiment indicator based on online search volume data has been constructed. We further introduce two new methods to estimate the rent component in a standard property yie ld model, as suggested in the literature (Chervachidze et al (2011); Sivitanides et al. (2001)). The first method is based on a quarterly forecast of annual rent changes. The second approach uses a twostep method, where the change of real rent is estimated first and expected changes of real GDP are used to reinforce the growth expectations of the market. Our panel data set ranges from 2004q1 2014q4 and comprises 80 regions within Europe. Our results show that a sentiment indicator can improve yield models in terms of forecast quality. Using an indicator based on online search volume data, shows that even online search behaviour is able to give insight about the mood of the market.