Hi there @nguyenhungsync, in order to help troubleshoot could you please let me know what sort of hardware you’re running on - that looks like an Orange Pi, is that right? And it’s running Ubuntu 16.04?
While I’m not familiar enough with the specifics here to say for sure, the fact that it’s looking for speech_recognition rather than SpeechRecongition may be part of the issue there? Even if it’s case-insensitive (which I don’t know one way or the other), that _ would still make a difference, I’d imagine.
Oh, that reminds me… I remember reading someplace that you can specify which Python you want pip to interact with when you have multiple versions installed by replacing pip in the command with pip2 or pip2.7 (apparently it varies which one you need to use between different OS variants? IIRC when I was trying (and failing) to install things in Cygwin I used pip2 quite a bit.) not sure how much that helps, but I figure it can’t hurt to try it if other things don’t work out first.