A method of hyperspectral image data compression using partitioned Karhunen-Loeve transform (PKLT) and integer wavelet transform (IWT) is proposed. IWT technique and set partitioning in hierarchical trees (SPIHT) compression coding are used to compress the PKLT data. In comparison with other methods, PKLT/IWT/SPIHT not only reserves the compression performance of KLT and the performance of IWT's high compression rate, but also is easy to implement real-time transmission. Experimental results show that the peak signal-to-noise ratio(PSNR) of PKLT/IWT/SPIHT improves 6 dB, 9 dB and 8 dB comparing with KLT/WT/WSFCVQ, KLT/WT/IBBZTC and KLT/WT/FSVQ in the condition of same compression rate respectively, computation hours decrease half, and the whole compression performance is enhanced much more.