Total variation $(TV)$ regularization method has been used in forward imaging of airborne scanning radar. However, the traditional regularization method usually processes the whole echo matrix line by line, and every processing includes inversion operation, which leads to huge computational complexity and insufficient real-time. To solve this problem, an improved TV regularization method is proposed for online real aperture radar super-resolution. By using the correlation between adjacent echoes, the matrix inversion operation is transformed into the iterative operation of matrix multiplication, and the real-time scanning echoes are used to effectively update the reconstruction results online, and the real-time online update of the target in the forward view area is realized. Compared with the traditional batch processing method, the proposed method significantly reduces the computational complexity without losing the imaging performance, saves the computing time, and is suitable for the fast imaging of scanning radar.