In 2009, Jeong et al. proposed a new searchable encryption scheme with keyword-recoverability which is secure even if the adversaries have any useful partial information about the keyword. They also proposed an extension scheme for multi-keywords. However, this paper demonstrates that Jeong et al.'s schemes are vulnerable to off-line keyword guessing attacks, where an adversary (insider/outsider) can retrieve information of certain keyword from any captured query message of the scheme.