In this study, wavelet analysis as a modern technique in financial and economic issues is used to evaluate capital asset pricing model. For this purpose, using market return, beta coefficient was calculated for daily return of 20 shares selected form Tehran Stock Exchange from March 2007 until March 2012. Then, the relationship between wavelet beta and mean stock return was evaluated on short-term, medium-term and long-term scales using least-squares technique. The findings indicate that the capital asset pricing model is a multi-scale model that uses different time scales for stock returns to provide different estimates of beta. Thus, Tehran Stock Exchange is a very efficient capital market on the long-term scale. Subsequently, values at risk of portfolios were calculated for each scale. According to the result, risk is highly concentrated on high-frequency time series (lower scales).