To realize the efficient use of battery residual energy, this paper attempts to estimate both the state of energy (SoE) and the state of available power (SoAP) for li-ion battery packs. First, the parameters of a 1st-order equivalent circuit model are identified online where the charging and discharging resistances are separately modeled. Then a state of energy estimator, considering the energy dissipation by heat convection, is designed using an unscented particle filter. Afterwards, multiple constraints in terms of cut-off voltages, recommended residual energy, extreme currents, and powers are incorporated to aid in SoAP prediction. Experiments on a 4-cell battery pack using a high-dynamic load profile show that the SoE estimator is reliable against various working conditions. The predicted SoAP with different time horizons and at different temperatures can avoid the conflicts with the preset constraints while giving reliable predictions.