A video encoding method comprises the steps of converting syntax data of an image block into a binary sequence; encoding each symbol of the binary sequence subject to its corresponding context; estimating the number of bits required for the binary arithmetic encoding of the syntax data of the image block; deciding whether to encode the syntax data of the image block to output it responding to the estimated bit number; and updating the context corresponding the syntax data decided to be encoded and to be outputted and not updating the context corresponding to the syntax data decided not to be encoded and not to be outputted. The estimating step uses the maximum value of the number of arithmetic encoding output bits for its input symbols as the aforementioned number of bits.