There are several difficulties with string theory: 1) it has made no falsifiable predictions. This is partly a consequence of the fact that it seeks to describe physics at energies far beyond anything realizable here on Earth. Because we have no experimental constraints on string theory, there are currently a vast, vast number of ways that the extra dimensions might arrange themselves (these ways are referred to as "solutions" of string theory). This problem has given rise to the "landscape" of string theory. 2) there are formidable theoretical stumbling blocks, including a great difficulty in getting inflationary spacetimes out of string theory, not to mention an apparent inability to recover the Standard Model of particle physics (or basic extensions, like MSSM) from string theoretic unifcation schemes.
The theoretical difficulties might be surmountable, but in my view the biggest issue is the lack of experimental support. In a very real sense, string theory is not only not a theory, it is not a science, in the sense that we generally use science to understand the world through observation and experiment. For now, string theory is an impressive theoretical formalism that proposes a universe built out of string; whether that universe happens to adhere to reality is still an open question.
Over all, my impression is that professional interest in string theory is waning.