We derive an adaptive power control method for a collaborative network utilizing partner selection that aims to minimize the frame error rate (FER). We model a decode-and-forward (DF) collaborative network under block fading conditions, which contains M independent users utilizing codes, whose performance can be expressed by a signal to noise (SNR) threshold, such as turbo codes. We show that partner selection can reduce system complexity and power allocation can improve the FER performance. We use both a search method as well as a convex deterministic method to demonstrate our power allocation scheme. This research extends other work in which adaptive power allocation is only applied to limited scenarios. We conclude that power control can greatly benefit a DF collaborative network in a block fading environment.