In the present study, the expected returns of cryptocurrencies were compared capital asset pricing model (CAPM) and downside capital asset pricing model (D-CAPM) approaches. For this purpose, fifty cryptocurrencies were studied as representative of risky assets in a five-year period from 2018 to 2022 with daily frequency. First, the panel was investigated using Levin-Lin-Chu, Im-Pesaran, and Shin and Dickey-Fuller's tests. Then, using paired t-statistics, the difference between the return estimates of the two models was investigated. Finally, using R2 and the generalized linear test model, the better model was selected to justify the changes in asset returns in these cryptocurrencies and portfolios. Based on the results, in almost 90% of the analyzed portfolios, the D-CAPM model was better than the CAPM model and had more justification power than the old CAPM model. In less than 1% of the models, the degree of justification and the appropriateness of the models were the same.