In this paper, the received signal in a range cell is modelled as a multi-component linear frequency-modulated (LFM) signal after range compression and motion compensation, and a new method based on segment integration and Lv's transform (LVT) is introduced for parameter estimation of LFM signals over long observation interval. In this method, the LFM signals are firstly divided into segments and fast Fourier transform (FFT) is then applied within each signal segment. After that, the same frequency resolution bins of each segment are selected to construct new series and inter-segment LVT is implemented to obtain the parameter estimates. The criteria to choose the number of segments, output signal-to-noise ratio, computational complexity and memory cost are analysed in detail for this new approach. This method is fast and able to obtain the accurate parameter estimates by using the complex multiplications and FFT. Comparisons with other popular methods, LVT, maximum-likelihood estimation and fractional Fourier transform are performed. Experimental results demonstrate the proposed method is capable of obtaining the accurate parameter estimates with low computational burden and storage memory, making it suitable to be applied in memory-limited and real-time processing systems.