Multimedia applications are ubiquitous in daily life. To satisfy multiple requests equipped with different devices and bandwidths, scalable video coding has been developed. However, it is difficult to realise spatial and quality scalabilities in practice owing to the increased coding complexity brought by inter-layer prediction. A novel fast mode decision algorithm for spatial scalable video coding is proposed in which macroblock modes in the base layer and the enhancement layer are predicted from each other. The experimental results show that the proposed algorithm provides up to 74% of time saving with ignorable quality degradation and acceptable bit rate increase.