A novel numeric method for the planar space manipulators joint trajectory optimization problem was proposed based on the wavelet collocation. Based on the interval spline wavelet according with its derivation algorithms, the approach expanded the state and control variables via cubic spline wavelet decomposition, and then the trajectory optimization problem was translated to nonlinear programming problem with wavelet coefficient being parameters. The wavelet advantageous properties of compact representation could inherently make it efficiently and accurately capture the discontinuities. The novel approach was demonstrated by a typical space manipulators joint trajectory optimization problem, and the optimal results show that out approach is efficient and accurate.