The concept of maternal affect attunement has evoked considerable theoretical interest, but attempts at empirical validation have been scarce. The aim of this study was to refine the coding scheme for assessment of maternal affect attunement and to establish the internal validity of the measure. Forty dyads with seven-month-old infants and their mothers were recorded in two face-to-face play situations. Mother-child interactions were coded by four raters on the dimensions of Maintaining Attention and Warm Sensitivity, the latter comprising the subscales of Positive Affect, Social Responsiveness, and Warm Concern. Inter-rater agreements for all scales and subscales were high. Scale reliabilities and construct stabilities as estimated by test-retest correlations were satisfactory. We conclude that the Maternal Affect Attunement Scale (MAAS) reliably captures individual differences in Maintaining Attention and Warm Sensitivity.