In this chapter, the gradient-based overdetermined/determined BSS algorithms are investigated. First, the principles of fixed step-size stochastic gradient BSS algorithms, natural gradient BSS algorithms, and EASI BSS algorithms are introduced in detail. On this basis, variable step-size BSS algorithms are investigated, including variable step-size natural gradient BSS algorithms and variable step-size EASI algorithms. and then, the stability conditions of the above algorithms as well as the selection of the nonlinear functions are derived in detail. Finally, the performance of algorithms presented in the chapter are analyzed by simulation.